GB2381872A

GB2381872A - Exhaust emission control system of an internal combustion engine

Info

Publication number: GB2381872A
Application number: GB0221913A
Authority: GB
Inventors: Marko Weirich
Original assignee: DaimlerChrysler AG
Current assignee: Mercedes Benz Group AG
Priority date: 2001-10-02
Filing date: 2002-09-20
Publication date: 2003-05-14
Anticipated expiration: 2022-09-20
Also published as: FR2830276A1; GB2381872B; DE10148663A1; GB0221913D0

Abstract

The level of the NO x emission from an internal combustion engine primarily varies as a function of the local temperature, the oxygen concentration and the length of time that the cylinder charge spends in the combustion chamber. These latter two variables are detected by measuring the engine speed, the air introduced and the quantity of fuel. It is particularly important in determining the untreated NO x emission, especially in the case of internal combustion engines operated with excess air and when using NO x storage catalytic converters, that for monitoring the NO x charge mass a different modelling be performed depending on the operating mode. Engine parameters such as fuel quantity, engine speed, exhaust gas recirculation rate, air ratio and ignition point, which are dependent on the respective operating mode are therefore registered and used along with the temperature of the combustion chamber to calculate the nitrogen oxide emission.

Description

- 1 Exhaust emission control system of an internal combustion engine The

invention relates to method of determining the nitrogen oxide content of the exhaust gases from internal combustion engines, in particular of spark ignition engines with direct fuel injection.

In the operation of internal combustion engines, exhaust gases are produced, which contain various pollutants, the respective proportions of these pollutants essentially varying as a function of the composition of the fueltair mixture. The proportion of nitrogen oxide emissions (NOx) is particularly great in the case of operation with a lean fueUair mixture, that is to say lambda >1. The use of NOX storage catalytic converters in the case of such engines, in order to be able to meet the stringent regulations on exhaust emissions, is already known. Despite a regeneration initiated in a specific sequence in operation of the engine, under certain operating conditions the NOx storage catalytic converters only have a limited storage capacity, so that adequate storage of the nitrogen oxides occurring is not always possible.

In discontinuous exhaust emission control processes it is necessary to determine the NOx mass reaching an NOx storage catalytic converter, in order to be able to determine the time and the duration of any regeneration. This NOx mass varies as a function of the NOx concentration in the exhaust gas and the exhaust gas mass flow and in operation with a lean mixture is essentially equal to the untreated NOx emission of the engine.

One possible method of determining the NOx concentration in the exhaust gas is to use a sensor upstream of the catalytic converter. This delivers a signal representing the NOx concentration in the exhaust gas, which can be used together with the air and mass fuel flow to calculate the NOx mass emitted. For diagnostic purposes such a sensor should also be arranged downstream of the catalytic converter, in order to monitor its efficiency.

Since the aforementioned sensors are expensive, it is desirable to be able to dispense with one or even both of the two sensors. Attempts are being made therefore to

- 2 - calculate the nitrogen oxide emission of the engine from corresponding contributory factors using a computer model. For this reason the influence of various engine parameters on the untreated NOx emission is incorporated into a structure for an untreated nitrogen oxide emission model, in order that, given sufficiently accurate modelling, at least one NOx sensor in the exhaust line may be dispensed with.

DE 19851319 Al discloses a method of determining the untreated NOx emission from an internal combustion engine operated with excess air, which is intended to approximate the modelled untreated NOx emission to the actual untreated NOx emission. In this method the degree of throttling and the inlet air temperature are quantitatively taken into account in a computer model for determining the untreated NOx emission. In this, a basic untreated NOx emission from the internal combustion engine in stratified-charge operation is obtained from a characteristic map as a function of the fuel mass and the engine speed. This basic value is then determined with a correction value varying as a function of the inlet air temperature, which value is in turn corrected by a correction value varying as a function of the engine speed and the exhaust gas recirculation rate. In addition a further correction is performed by taking into account the degree of throttling in stratified-charge operation.

The aforementioned method fails to take account of some engine parameters that have a considerable influence on the NOx formation in the engine.

The present invention therefore seeks to create a method of determining the nitrogen oxide content of the exhaust gases from internal combustion engines, by means of which the Oxygen oxide emission can be determined more accurately.

According to a first aspect of the present invention there is provided a method of determining the nitrogen oxide emissions contained in the exhaust gas from an internal combustion engine, having a combustion chamber arranged in a cylinder, a catalytic converter, and exhaust gas recirculation device and an engine control, - a fuel quantity delivered to the combustion chamber, - an air quantity delivered to the combustion chamber, - an exhaust gas recirculation rate, and - a speed of the internal combustion engine being registered, __ i a ellelllililall_ ellaa lllilll If l if Ill l Alp-Ill II I l In l lit ll l llllllIIII IIIIILL IIII III I 11 i l li It fillip_

wherein - the thermal condition of the combustion chamber is registered in stratified-

charge operation and - the mass of the nitrogen oxide emission is calculated from a value for the thermal condition of the combustion chamber together with values for the fuel quantity, the engine speed and the exhaust gas recirculation rate.

According to a second aspect of the present invention there is provided a method of determining the nitrogen oxide emissions contained in the exhaust gas from an internal combustion engine, having a combustion chamber arranged in a cylinder, a catalytic converter, and exhaust gas recirculation device and an engine control, - a fuel quantity delivered to the combustion chamber, - an air ratio through an air quantity delivered to the combustion chamber, - an exhaust gas recirculation rate, - an ignition point, and - a speed of the internal combustion engine being registered, wherein - the thermal condition of the combustion chamber is registered in homogeneous operation and the mass of the nitrogen oxide emission is calculated from a value for the thermal condition of the combustion chamber together with values for the fuel quantity, the engine speed, the air ratio and the ignition point.

The present invention allows the NOx emission to be accurately calculated, since this calculation is based on the actual factors contributing to the NOx emission. The level of the NOx emission from an internal combustion engine primarily depends on the local temperature, the oxygen concentration and the length of time which the cylinder charge spends in the combustion chamber. The latter two variables can be determined relatively easily by measuring the engine speed, the air introduced and the quantity of fuel. According to the present invention, it is particularly important in determining the untreated NOx emission, especially in the case of internal combustion engines operated with excess air and when using NOx storage catalytic converters, that for monitoring the NOx charge mass a different modelling be performed depending on the operating mode. The aforementioned internal combustion engines are preferably run

- 4 - in stratified harge operation or homogeneous lean-burn operation. Some engine parameters are therefore registered as a function of the respective operating mode and used to calculate the nitrogen oxide emission. At the same time, for a precise determination or approximation of the untreated NOx emissions, the corresponding signals are registered by a control unit and the NOx emissions then determined by modelling according to the invention.

The method according to the invention for determining the nitrogen oxide emissions provides for sensing of the quantity of fuel delivered to the combustion chamber, the quantity of air delivered to the combustion chamber, the exhaust gas recirculation rate, the speed of the internal combustion engine, the ignition point and the thermal condition of the combustion chamber, the mass of the nitrogen oxide emission in stratifiedcharge operation being calculated from the value for the thermal condition of the combustion chamber and the values for the fuel quantity, the speed and the exhaust gas recirculation rate. The mass of the nitrogen oxide emission in homogenous lean-

burn operation, on the other hand, is calculated from the value for the thermal condition of the combustion chamber and the values for the fuel quantity, the speed, the air ratio and the ignition point.

According to the invention the thermal condition of the combustion chamber is detected by determining the combustion chamber temperature, which can be done by means of a temperature sensor. In order to detect the combustion chamber temperature level, the temperature sensor preferably measures the cylinder wall temperature in the combustion chamber area or the cylinder head temperature in the combustion chamber area. According to a preferred embodiment of the invention, the thermal condition of the combustion chamber can be determined by registering the coolant temperature, particularly in the area of the combustion chamber. That is to say the coolant temperature of the engine is determined in the cylinder head above the combustion chamber or on the cylinder wall side in the area where ignition of the fuel/air mixture occurs.

The untreated NOx emissions in stratified-charge operation are determined taking into account the thermal level of the combustion chamber, the quantity of fuel involved in the cornhustion, the engine speed and the exhaust gas recirculation (EGR) rate.

According to a basic characteristic map filed in the control unit, the quantity of fuel and the speed of an internal combustion engine characterize the current operating 11. 1 ' - 11 1 1 1111' 11 18 1!11515, !11 lit' 1 - '1 'em 191 11111111111 1011

- 5 - point. Such a characteristic map gives an NOx value, which is first corrected by an EGR rate correction value. This is in turn corrected by a value determined from the thermal condition of the combustion chamber of the internal combustion engine, in order to determine the actual NOx value.

The quantity of fuel may be determined, for example, by a fuel measuring device or through values such as the fuel injection period and the pressure on the fuel injector.

The engine speed can also be registered by means of a speed sensor. The exhaust gas recirculation rate is determined at the exhaust gas recirculation valve by means of a suitable measuring device, it being possible to detect the mass air flow from a throttle valve device.

The correction values are then determined by means of various characteristic maps filed in the control unit and are used to calculate the actual untreated NOx emission on the basis ofthe modelling according to the invention.

The untreated NOx emissions are determined in homogeneous operation, particularly in a homogeneous lean-burn operation (air ratio approx. 1.4 to 1.5), taking into account the thermal level of the combustion chamber, the quantity of fuel involved in the combustion, the engine speed, the air ratio and the ignition point. As in stratified-charge operation, a correction value, an ignition point correction value and a combustion chamber thermal condition correction value are determined, which are used to calculate the untreated NOx emission on the basis of a modelling according to the invention.

According to preferred embodiments of the invention, a 2 order polynomial is used in calculating the mass of the nitrogen oxide emission for an approximation of the relationships between the air ratio, the exhaust gas recirculation rate and the nitrogen oxide emission. In calculating the mass of the nitrogen oxide emission the order of the input variables is freely selectable and the input variables can be combined either by multiplication or addition.

According to a further preferred embodiment of the invention the calculated mass of the nitrogen oxide emission can be corrected through the air ratio for the individual cylinders of the internal combustion engine. This makes it possible to calculate the untreated NOx emission of an internal combustion engine precisely, despite

differences in combustion air ratio between the individual cylinders. Since the air ratio in the individual cylinders is not always the same, owing to variations in the manufacture of fuel injectors, corresponding differences in the NOx emissions can also occur. A lambda closed-loop control usually balances out such differences by supplying all cylinders with more or less fuel, so that the overall air/fuel ratio required in the exhaust gas is set. Since the NOx emission from a cylinder does not vary as a linear function of the air ratio, however, this means that the NOx emission calculated by an untreated NOx emission model does not match the measurable NOx emission. It is nevertheless possible, by determining the values for the individual cylinders, to detect the differences in the fuel injection quantities. This can be achieved, for example, by means of a high time resolution of the signal measured by a control probe. The firing order of the internal combustion engine can be identified from the exhaust gas, in such a way that the probe registers the deviations of the individual cylinders from the set-point air ratio. The NOx emission of the engine it therefore determined as the sum of the emissions from the individual cylinders.

It is moreover advisable to provide a NOx sensor, which detects the NOx fraction in the exhaust gas flow, the resulting measured value being compared with the value for the calculated NOx emission.

Further advantages and details of the invention are contained in the claims and in the description. The invention is explained in more detail below with reference to the

drawing, in which: Fig. I shows a diagram of a spark-ignition engine with direct fuel infection having exhaust gas recirculation and an NOx storage catalytic converter, Fig. 2 shows the influence of the exhaust gas recirculation rate on the NOX concentration in the exhaust gas in stratified-charge operation, Fig. 3 shows the influence of the air ratio on the NOx concentration in the exhaust gas in stratified-charge operation, Fig. 4 shows the influence of the coolant temperature on the NOx concentration in the exhaust gas in stratified-charge operation, I,,,, ,q,.,,,_, Ale, $ Aim 1 _ 81 D I I 111' 1 1 11111 111'! 1 1111111 1111 i I i 11 1 11 18

Fig. 5 shows the influence of the air ratio on the NOx concentration in the exhaust gas in homogeneous operation, Fig. 6 shows the influence of the ignition point on the NOx concentration in the exhaust gas in homogeneous operation, Fig. 7 shows the influence of the coolant temperature on the NOX concentration in the exhaust gas in homogeneous operation, Fig. 8 shows the principle of a combination through addition.

Fig. 9 shows the principle of a combination through multiplication, Fig. 10 shows how the influence of load and engine speed are taken into account in the NOx correction, Fig. 1] shows a calculation of a 2n order polynomial in the engine control unit, Fig. 12 shows a model structure for determining the NOx emission in stratified-

charge operation, Fig. 13 shows a model structure for determining the NOx emission in homogeneous operation, Fig. 14 shows a diagram of a cylinder block of a four cylinder internal combustion engine with lambda probe fitted, Fig.] 5 shows a graph of the NOx emission as a function of the air/fuel ratio, and Fig.]6 shows a curve for the signal of a four cylinder internal combustion engine. Figure 1 represents a cylinder block] of a 4-cylinder spark ignition engine with direct fuel injection showing a diagrammatic cross-section through a cylinder. A fuel injector 2 and a spark plug 3 are arranged in each cylinder. The cylinder also contains a combustion chamber 4, which is defined by a displaceable guided piston 7. Fresh l

air is delivered to the combustion chamber 4 through an inlet valve 5, the exhaust gases following combustion being discharged though an exhaust valve 6.

A control unit 8 controls the combustion and the processing of the various signals.

The control unit 8 registers the signals from a lambda probe 9, an NOx storage catalytic converter 10, an exhaust gas recirculation valve 11, a temperature sensor 12 arranged in the area of the combustion chamber 4, a throttle valve device 13 and a Fuel measuring device 17 arranged downstream of an injection pump 16. The signal processing in the control unit 8 is an integral part of an electronic circuit (not shown).

A disc (not shown), which may simultaneously form the flywheel, for example, is arranged on a crankshaft, not shown in the drawing, of the internal combustion engine, an angle mark sensor (not shown) being assigned to the said disc for registering the engine speed. This angle mark sensor is also connected to the control unit 8.

The combustion chamber 4 is defined at its upper end by a cylinder head IS, the inlet and exhaust valves being arranged in the cylinder head IS. Through the inlet valve 5, the necessary combustion air can flow into the combustion chamber 4 through the air inlet pipe 19, the prevailing air mass in the throttle valve device being registered by an mass air flow sensor, not shown. The mass air flow sensor is likewise connected to the control unit 8, it being feasible to determine the air mass from the throttle valve position and the engine speed.

The exhaust gases pass through the exhaust valve 6 into an exhaust line 21, which leads to the NOx storage catalytic converter 10. An exhaust gas recirculation line 14 branching off Mom the exhaust line 21 is provided, in order to return exhaust gas into the inlet pipe 19. In this exhaust gas recirculation line 14 a return flow sensor (not shown) is arranged in the exhaust gas recirculation valve 11, the sensor registering the mass of the recirculated exhaust gas and transmitting corresponding signals to the control unit 8.

The calculation of the NOx mass flow from the NOx concentration in the exhaust gas c (NOX) is explained in more detail below. The NOX mass flow moo, is calculated from the NOX concentration c(NO,) in the exhaust gas according to the following fonnula: .' _ - 1ll ye 1llil 1l 1l! | ll I 1 l ill l l l | 1158 I l All I 1ll l

m,40 = M(NO2 KNOX FA (at At) The molar mass M of NO2 and the exhaust gas is used for evaluation of the nitrogen oxide emissions. FW and FA serve to correct the atmospheric humidity, i.e. the NOx emission is scaled to a specific atmospheric humidity, MBH representing the mass air flow and MLKH the mass fuel flow. This is necessary because the atmospheric humidity has a measurable influence on the NOx emission of internal combustion engines. A lower NOx emission is produced at a higher atmospheric humidity than in dry air. The heat capacity of the intake air increases with the water vapour content, thereby reducing the combustion peak temperatures, FW being calculated according to the following formula: FW" 1 1+a (7 PHIT75)+b 1,8 ( ANSK-3023 where: PHIT: is the atmospheric humidity in relation to dry air in g/kg XTANSK: is the intake air temperature in K The factors a and b are determined according to the following formulae: MLKH-PDV '

b - 51 46 1 1-LKH-PDV +0'0053

- 1 0 The partial pressure ratio PDV may be calculated via the dew point temperature Ttau (in C): PDV XPL - PDT

where: XPL: is the absolute air pressure PDT: is the vapour pressure calculated as: PDT = 3,267.107 Ttau4 + 0,0002988 Ttau3 + 0,01408 - Ttau2 + 0,4378 Ttau + 6,054 The correction factor PA is obtained from the term FA-1-1 5 M H PDV

The influence of the atmospheric humidity cannot be taken into account when calculating the NOX mass flow in the vehicle, since no corresponding sensors are available for determining atmospheric humidity or dew point. For this reason a simplified calculation must be used, it being possible to take the influence of the atmospheric humidity into account if such sensors do happen to be fitted in the vehicle, or corresponding information is available via data communications. The simplified calculation only takes account of the exhaust gas mass flow, which is composed of the mass air flow MBH and mass fuel flow MLKH.

mhiox = M g, Knox) (MSH MLKH) .I_,. ''ll ,., _lel: I IllillI 18 D! Illill'l II 111 1111 1l 1 It ill 1 llllI 1 1llllllllI'l.

- 11 The NOx emission of an engine fundamentally varies as a function of engine speed and load. In the case of a spark-ignition engine with direct fuel injection, the load is preferably adjusted by way of the fuel mass delivered (MFF). This variable is determined by the fuel measuring device 17 and is relayed by corresponding signals to the engine control unit 8.

The fuel mass delivered can therefore be used to determine the NOx emission.

The storage of characteristic maps, which for stratified-charge operation represent the nitrogen oxide emissions for the engine used as a function of the load points run, is of particular importance to the present invention. Such a characteristic map is referred to as "basic characteristic map". If all boundary conditions in the vehicle correspond to those on an engine test bench used to determine such characteristic maps, an NOx value is obtained which is stored in the characteristic map. In dynamic operation, however, certain parameters differ from the values measured when stationary. Any deviation in the parameters results in an adjustment of the quantity of NOx emitted.

The influence of significant parameters on the NOx emission is explained below.

The applied exhaust gas recirculation rate has the greatest influence on the NOx emission. The exhaust gas recirculation (EGR) is used to reduce the NOx emission.

Figure 2 shows the curve for the NOx concentration given a variation in the exhaust gas recirculation rate (EGR rate) in stratified-charge operation. The NOX concentration falls as the exhaust gas recirculation rate rises. A deviation of the EGR rate from the set-point value in dynamic operation results, in a first approximation, in a linear decrease or increase in the NOX emission. The gradient of the straight lines varies as a function of the load, characterized by the injected fuel quantity (MFF), and engine speed (n). The exhaust gas recirculation rate (EGR) is not measured directly on the vehicle, but is calculated, for example, via a lift of the EGR valve 11 and a pressure differential, calculated from an inlet pipe model, between the pressure in the inlet pipe 19 and the pressure in the exhaust line 21. The quality of an NOx correction for the exhaust gas recirculation is therefore heavily dependent upon the quality of the determination of the EGR rate on the vehicle.

Another parameter having an influence on the NOx emission in stratifiedcharge operation is the air ratio if). Figure 3 shows the curve for the NOX concentration as a

- 12- function of In dynamic operation the actual air ratio may deviate from the value in stationary operation, if the throttle valve is not set to the applied value rapidly enough, for example. The air deficiency or excess has an effect on the combustion and hence on the NOX emission. The air ratio is set for the respective measurements by adjustment of the throttle valve position. With exhaust gas recirculation activated, the EGR rate would also vary significantly due to the changing negative pressure in the inlet pipe l9 and would have an influence on the NOx emission. The curves represented in Figure 3, therefore were determined without exhaust gas recirculation.

The relationship is not linear but can be approximated by a 2nd order polynomial.

According to Figure 3 the coefficients of such a polynomial differ according to load MFF and engine speed n.

The temperature of the combustion chamber 4 also has an influence on the combustion temperature and hence on the NOx emission. According to a preferred embodiment of the invention the coolant temperature may be used as an indicator of the combustion chamber temperature, although no direct correlation exists in warm-up operation of the engine. Figure 4 shows the NOX concentration in the exhaust gas as a function of the coolant temperatures from 40 to 90 C for stratified-charge operation, the values shown in Figure 4 being obtained at a constant EGR rate of 20%. Altematively, the combustion chamber wall temperature can, according to the invention, serve as an indicator of the combustion chamber temperature, which is measured by the temperature sensor 12. A cold combustion chamber wall reduces a combustion peak temperature and hence the NOX emission. On the other hand, mixture formation is improved in a hot combustion chamber, which leads to an increased NOX emission.

The relationship between coolant temperature and NOX emission lends itself to linear approximation. The gradient of the straight lines according to Figure 4 is virtually independent of the load MFF and engine speed n.

As in stratified-charge operation, the NOX emission in homogeneous operation also varies as a function of the load and engine speed. In a stoichiometric operation a three-

way catalytic converter is generally capable of reducing the nitrogen oxides. In homogeneous lean-burn operation, on the other hand, as in the upper load range of a modem spark-ignition engine with direct fuel ignition, for example, knowledge of the NOX emission is needed in order to calculate the NOX mass stored in the NOX storage catalytic converter lo.

.,,,,,,.,, 111-IX 1 11! ' IF'I R15 1! ' 8 ' ' I!l I 1 1. IF 11 1 l 11P'!! 11 11 I1111111!11111!

- 13- In homogeneous operation the load is controlled by way of the charge or air mass delivered. The value for the air intake mass is calculated in the inlet pipe model of the engine control unit and can also be measured by a hot-film mass air flow sensor, not shown in Figure l, for adjustment of the model. In order to obtain the maximum fuel economy, the set-point selected for the air ratio may be as lean as possible, that is to say just short of a running limit for operation without misfiring, where the air ratio value (a) is approximately l.4 to 1.5. In such a range it is no longer usual to use exhaust gas recirculation in order to reduce the NOx emission, since this has an excessively adverse effect on the smooth running of the engine. Figure 5 shows the nitrogen oxide emission for an homogeneous lean-burn operation (\ is approximately 1.4) as a function of the air ratio with the engine speed n and the mass air flow MAF as load parameters. It will be seen that the nitrogen oxide emission in homogeneous lean-burn operation is significantly higher than that in stratified-charge operation.

In homogeneous lean-burn operation, the air ratio has the greatest influence on the NOx formation. Figure 5 shows such a relationship with a constant ignition point. It must be noted that the relationship is nonlinear, although for minor fluctuations about a set-point air ratio this lends itself to linear approximation. The gradient of a regression line is dependent upon the load MAF and the engine speed n.

In homogeneous lean-burn operation, it may be necessary to set a different ignition advance angle IGA from that applied in the set characteristic map. Reasons for this include knock-control, for example, or interventions to improve the ride comfort by avoiding torque surges. Such intervention in the ignition advance angle results in a deviation of the NOx emission from the basic value, as shown in Figure 6. Given a constant air ratio, the variation in the NOX emission is virtually a linear function of the ignition point IGA. Here the gradient of the regression lines differs according to load MAF and engine speed n.

In homogeneous lean-burn operation, too, the combustion chamber temperature has an influence on the NOx emission of an engine. In cold ambient conditions the peak temperature of the combustion is equally reduced. In homogeneous operation too, mixture preparation in a sparkignition engine with direct fuel injection, for example, is also inferior in a cold combustion chamber. In the case of a lean air ratio this may even lead to misfiring. Figure 7 represents the influence of the coolant temperature, identified by TCO, as indicator of the combustion chamber temperature, on the NOx

- 14 emission. According to Figure 7, the relationship between a coolant temperature and the NOx emission can be represented as a linear one. In a first approximation, the influence of the engine speed on the gradient of a regression line is negligible. If the relationships previously described are simulated in a model, it

must be assumed that individual parameters can be superimposed independently of one another. The basic characteristic map in the corresponding operating range in each case forms the basis for the calculation of an emitted NOX mass. The BOX value may be given as the concentration [ppm] or mass flow [mg/s]. In both cases the inaccuracy lies in the atmospheric humidity in the vehicle that is not taken into account. Giving the value in mg/s saves calculation time in the control unit 8, since there is no need to multiply by the air and fuel mass.

The value for the basic characteristic map can be corrected by factors or addends, which are obtained from the deviations of the aforementioned parameters from the set-

point value. The straight line relationship between a parameter and the NOx emission represents the simplest relationship as follows: NOx=a.x+b There are two possible ways of determining a corrected NOx value. The combination is performed either by addition or multiplication. Figure shows the principle of additive combination. According to this, the difference in the quantity of NOx as a function of a parameter deviation (x_rnes - x_ref) is stored in a correction characteristic map. If Ore is no deviation, the correction addend is set to zero.

Figure 9 shows the principle of multiplicative combination. The NOx emission relative to the reference value is now filed non-dimensionally in the correction characteristic map and as factor is multiplied by an NOx value of the basic characteristic map.

The difference between additive and multiplicative combination is that in a multiplicative combination the corrected NOx quantity can never be less than zero.

The product of reference valve and correction can therefore never be negative. On the other hand, it is possible in the case of additive combination, for the amount of a calculated correction addend to be greater than the reference value of the NOx ...,._,,__.,, 11 11 r 1111 1 111_ I 111 1 1 B1111li l1 Illi I

-15- emission. If the correction addend has a minus sign, the corrected NOx value may be less than zero. The additive combination is used for preference, since with a subsequent limiting it requires less calculation time in the processor of the engine control unit than the multiplication by a decimal number. Since the influence of a parameter may vary according to load and engine speed, this being expressed in a different gradient of the regression curve, the structure represented in Figure l0 is preferably used in order to take account of this effect.

The factor or the addend for the NOx correction is no longer stored in the correction characteristic map, but the gradient of the regression curves. This is multiplied by the parameter deviation, giving the correction value.

If the relationship between a parameter and the NOx emission cannot be described by a straight line, the accuracy must be improved by an Order polynomial. However, complex mathematical functions cannot be calculated in the processor of an engine control unit. Only additions and multiplications are possible.

One possible way of representing a 2 order polynomial, for example, by successively performing straight-forward additions and multiplications is described below.

NOx = a. x2 + b. x + c and: NOx_ref = a. x_ref 2 + b. x ref + c Here "x" is the actual value of the parameter currently measured, "x_ref" is the set-

point value of the parameter for the corresponding operating point, "a", "b" and "c" are coefficients dependent upon the corresponding operating point. The actual NOx emission is to be calculated. The addend "c" is removed by subtraction and the following equation is obtained: NOx NOx_ref = a. (Ax) 2 +b. (Ax) The coefficients "a" and "b" can be filed in characteristics curves. The structure then looks like that shown in Figure l l. Since an error in determining the coefficients is exponentiated by the repeated multiplication, a linear approximation is to be preferred if possible, even if the regression line does not represent the curve of the measure values exactly.

-16- The structures described above therefore permit the description of linear or polynomial

functions of up to 3 dependent parameters on the NOx emission. Further parameters having a separable influence can also be corrected.

The structure represented in Figure 12 is proposed for stratified-charge operation, owing to the different measurements.

The NOx quantity in mg/s emitted when stationary is filed in a basic characteristic map as a function of load MFF and engine speed n. This value is corrected in the event of differences between an adjusted EGR rate and a set-point value. The gradient of an approximation line, which describes the relationship between EGR rate and NOx mass flow, is filed in an EGR correction characteristic map. Multiplied by the difference between the actual and the set-point EGR rates, this gives an NOx mass flow, which is added to the value of the basic characteristic map. Finally the calculated NOx emission is correspondingly adjusted to the thermal condition of the engine by a factor varying as a function of the coolant temperature TCO.

According to the invention the model described in Figure 13 is proposed in order to calculate the NOx emissions in a homogeneous lean-burn operation. The NOx emission in mg/s as a function of the mass air flow MAE and engine speed is filed in a basic characteristic map. A correction is then performed if the actual air ratio does not correspond to the set- point air ratio. In the case of an intervention in the ignition advance angle IGA, the NOx quantity is likewise increased or reduced by an addend.

Finally, as Hi stratified-charge operation, a correction is performed by way of the coolant temperature TCO according to the thermal condition of the engine.

Owing to variations in the manufacture of the fuel injectors, it is possible that individual cylinders will not receive the allotted fuel quantity that is necessary for a certain set-point air/fuel ratio. As a result, it is possible that the individual cylinders will be operated with different air/fuel ratio, as is shown, for example, in Figure 14.

Although the lambda control balances out such tolerances by supplying all cylinders with more or less fuel, so that the overall air/fuel ratio required in the exhaust gas is set, the NOx emission of a cylinder is nevertheless not a linear function of the air ratio, as can be seen from Figure 15. This means that the Box emission calculated by an untreated NOx emission model will not correspond to the measurable NOx emission.

.,,._._ _,._, Ill 111 B 11111 18 1- 1151818 1 1 Al 1' 1 11 1 t1111 _IF10111111 1 1 111111' 11 115181 11 1 1111111111 1 1 11111111 11511

-17- By measuring the values of the individual cylinders, it is possible to detect the differences in the fuel injection quantities. This can be achieved through a high time resolution of the signal measured by the control probe, for example. The firing order of the internal combustion engine can be identified from the exhaust gas, as is shown for a "cylinder engine in Figure 16. If the engine is operated in accordance with Figure 15, the probe registers the deviations of the individual cylinders from the set-

point air ratio. The NOx emission of the engine is therefore obtained from the sum of the emissions from the individual cylinders, as follows: NOx,g's = Noxn + Nox.2 + NOx.3 + NOx.4 The NOx emission from an individual cylinder is obtained from the basic value and a correction on the basis of the difference between the basic value and the measured value. NOx = NOx, teas + ANOx, where Knox = dN;.

It is therefore possible to determine the untreated NOx emission of an internal combustion engine accurately despite differences in the combustion air ratio between individual cylinders.

By means of the present invention it is possible to monitor the NOx emission mass. It is furthermore possible to carry out monitoring of the uniformity of combustion in the individual cylinders. The use of an additional NOx sensor for system redundancy is also conceivable, in order to compare the measured value with the calculated value for NOx. The values determined for NOx may be used to control or regulate exhaust emission control systems. The present invention is particularly suited to use in the vehicle for so-called on-board diagnosis both in applied ignition and in compression ignition internal combustion engines, and permits constant calculation and monitoring of the NOX emission for a 2stroke or 4-stroke internal combustion engines. It is also feasible to transfer the findings of the present invention for performing tests on test benches.

Claims

1. A method of determining the nitrogen oxide emissions contained in the exhaust gas from an internal combustion engine, having a combustion chamber arranged in a cylinder, a catalytic converter, and exhaust gas recirculation device and an engine control, - a fuel quantity delivered to the combustion chamber, - an air quantity delivered to the combustion chamber, - an exhaust gas recirculation rate, and - a speed of the internal combustion engine being registered, wherein - the thermal condition of the combustion chamber is registered in stratif ed-

2. A method of determining the nitrogen oxide emissions contained in the exhaust gas from an internal combustion engine, having a combustion chamber arranged in a cylinder, a catalytic converter, and exhaust gas recirculation device and an engine control, - a fuel quantity delivered to the combustion chamber, - an air ratio through an air quantity delivered to the combustion chamber, - exhaust gas recirculation rate, - an ignition point, and - a speed ofthe internal combustion engine being registered, wherein - the thermal condition of the combustion chamber is registered in homogeneous operation and - the mass of the nitrogen oxide emission is calculated from a value for the thermal condition of the combustion chamber together with values for the fuel quantity, the engine speed, the air ratio and the ignition point.

_, ''_., r5 r ' _ r ''t.. _ t ' I! 15

- 1 9 A method according to Claim I or 2, wherein the combustion chamber temperature is registered in order to determine the thermal condition of the combustion chamber.

4. A method according to Claim 3, wherein the combustion chamber temperature is registered by detecting a cylinder wall temperature in the combustion chamber area or a cylinder head temperature in the combustion chamber area.

S. A method according to Claim 3, wherein the combustion chamber temperature is registered by detecting a coolant temperature by means of a temperature sensor, in the area of the combustion chamber.

6. A method according to any one of Claims I to 5, wherein a 2n order polynomial is used in calculating the mass of the nitrogen oxide emission as an approximation of the relationship between the air ratio, the exhaust gas recirculation rate and the nitrogen oxide emission.

7. A method according to any one of Claims 1 to 6, wherein in calculating the mass of the nitrogen oxide emission the input variables are in any order and the input variables are combined by multiplication.

8. A method according to any one of Claims 1 to 6, wherein in calculating the mass of the nitrogen oxide emission the input variables are in any order and the input variables are combined by addition.

9. A method according to any one of Claims I to 8, wherein the calculated mass of the nitrogen oxide emission is corrected by the air ratio for the individual cylinders of the internal combustion engine.

10. A method according to any one of Claims I to 9, wherein the air ratio of an individual cylinder of the internal combustion engine is registered by a time resolution of a signal measured by a lambda control probe.

11. A method according to any one of the preceding claims wherein the engine operates in lean burn mode with excess air.

- 20

12. A method of determining the nitrogen oxide emissions contained in the exhaust gas from an internal combustion engine, substantially as described herein with reference to and as illustrated in the accompanying drawings.

,,, _.,,,.,.,,,,,,,,,, I,.,.,,._. -,.,, _. _