WO2015059034A1

WO2015059034A1 - System and method for estimating the flow of nitrogen oxides in the exhaust gases of an internal combustion engine for a motor vehicle

Info

Publication number: WO2015059034A1
Application number: PCT/EP2014/072265
Authority: WO
Inventors: Felipe CASTILLO; Gilles MAUVIOT; Vincent Talon
Original assignee: Renault S.A.S.
Priority date: 2013-10-24
Filing date: 2014-10-16
Publication date: 2015-04-30
Also published as: FR3012526B1; CN105683547A; EP3060784A1; FR3012526A1

Abstract

The invention relates to a system for estimating the flow of nitrogen oxides in the exhaust gases of an internal combustion engine for a motor vehicle provided with at least one circuit for partially recirculating the exhaust gases and with at least one system for reprocessing the exhaust gases, the flow of nitrogen oxides being estimated upstream of the systems for reprocessing the exhaust gases, including a system for determining the mass fraction of the fresh gases, a subtractor being suitable for determining the fraction of combusted gases in the intake collector from the mass fraction of the fresh gases, and an estimator (1) of the flow of nitrogen oxides depending on the speed of rotation of the internal combustion engine and on the fraction of combusted gases in the intake collector.

Description

System and method for estimating the nitrogen oxide flow rate in the exhaust gas of an internal combustion engine for a motor vehicle.

The technical field of the invention is the control of an internal combustion engine, and more specifically the determination of the flow rate of nitrogen oxides in the exhaust gas of a motor thus controlled.

The pollution control standards require manufacturers to design more and more efficient engines equipped with increasingly advanced exhaust gas treatment systems. Today diesel engines for the Euro 6 standards are all equipped with nitrogen oxides catalyst, denoted NOx thereafter, or selective catalytic reduction (acronym SCR, for "Selective Catalytic Reduction"). For the proper functioning of these post-treatment systems, it is important to know the NOx loading of these systems in order to optimize the purge phases just needed. Indeed, in the case of nitrogen oxide catalysts (also known as NOx-Trap), the purges lead to an overconsumption of fuel and therefore an increase in C0 ₂ emissions. In the case of SCRs, the purges cause a consumption of urea.

Under these conditions, the manufacturers use NOx nitrogen oxide sensors upstream of these exhaust gas treatment systems in order to measure their flow rates. Today, in order not to have to bear the additional cost of this sensor some manufacturers reconstruct this information from pressure measurements in the cylinders. Documents FR2922262, FR2936015 and FR2945320 illustrate this state of the art.

However, this solution of NOx nitrogen oxides flow rate estimation with cylinder pressure measurement is very efficient, but requires a particular instrumentation of the engine with less a cylinder pressure sensor and a computer more efficient than the average.

The sensors and the computer required for the implementation of this technique generate a significant additional cost while presenting a certain measurement uncertainty.

Such an additional cost is incompatible with the reduction of costs incurred for entry and mid-range vehicles.

There is therefore a need for an estimation of the flow rate of nitrogen oxides in the exhaust gases of an internal combustion engine estimated upstream of exhaust gas reprocessing systems, especially when the engine is equipped with a partial recirculation circuit of the exhaust gas.

An object of the invention is a system for estimating the flow of nitrogen oxides in the exhaust gases of an internal combustion engine for a motor vehicle equipped with at least one partial recirculation circuit for the combustion gases. exhaust system and at least one exhaust gas reprocessing system, the nitrogen oxide flow rate being estimated upstream of the exhaust gas reprocessing systems. The system comprises a system for determining the mass fraction of fresh gas, a subtractor able to determine the fraction of gases burnt in the intake manifold from the mass fraction of fresh gas, and an oxide flow estimator. of nitrogen as a function of the rotational speed of the internal combustion engine and the fraction of burnt gases in the intake manifold.

The oxide flow rate estimator may include a mapping of the nitrogen oxide flow rate inputted to the rotational speed of the internal combustion engine and the fuel flow, as well as a shape function correction means. as a function of the fraction of gases burnt in the intake manifold, and a multiplier connected at the input to the outputs of the mapping and the correction means, able to output the estimation of the flow of nitrogen oxides. of the internal combustion engine.

The nitrogen oxide flow rate estimator may include an algebraic relationship estimation method that is a function of the richness, fuel flow, the fraction of flue gases in the intake manifold, and the phase of the main injection.

Another object of the invention is a method for estimating the flow of nitrogen oxides in the exhaust gas of an internal combustion engine for a motor vehicle equipped with at least one partial recirculation circuit of the combustion gases. exhaust and at least one exhaust gas reprocessing system, the flow of nitrogen oxides being estimated upstream of exhaust gas reprocessing systems. The method comprises the following steps:

a value of the mass fraction of fresh gas is estimated, the value of the fraction of gases burnt in the intake manifold is determined as a function of the mass fraction of fresh gas, and

a value of the flow of nitrogen oxides is determined as a function of the fraction of gases burned in the intake manifold and the fuel flow rate.

The rate of nitrogen oxides can be determined as a function of engine speed and fuel flow mapping, and the nitrogen oxide flow values can be corrected for a range of times. function of form depending on the fraction of gases burned in the intake manifold.

The flow rate of nitrogen oxides can be determined as a function of an algebraic relationship and the fuel flow, the richness, the phasing of the main injection and the fraction of burned gases.

From measurements of nitrogen oxide flow rate at constant engine rotation speed, at constant torque, and with partial recirculation of the exhaust gas, the flow rate of nitrogen oxides can be determined for operating points corresponding to the operating points for which the flow rate of nitrogen oxides has previously been determined without partial recirculation of the exhaust gases, it is possible to memorize the ratio between the flow rate of nitrogen oxides without partial recirculation of the exhaust gases. and the flow of nitrogen oxides with partial recirculation of the exhaust gases for a set of values of the fraction of flue gases in the intake manifold, and

the function of form can be determined as the average function of all these points.

Other objects, features and advantages of the invention will become apparent on reading the following description, given solely by way of nonlimiting example and with reference to the appended drawings in which:

FIG. 1 illustrates the principal elements of a first embodiment of a nitrogen oxide flow estimator according to the invention,

FIG. 2 illustrates the main elements of an internal combustion engine equipped with a double exhaust gas recirculation circuit,

FIG. 3 illustrates the main elements of a system for determining the mass fraction of fresh gas in the intake manifold, and

FIG. 4 illustrates the main elements of a second embodiment of a nitrogen oxide flow estimator according to the invention.

The NOx nitrogen oxide flow rate estimator 1 should be able to provide a +/- 20% accuracy reading recommended for use on the estimation of NOx loading of post - treatment systems. Such an NOx nitrogen oxide flow rate estimator may be designed to require only the engine rotational speed (in rpm), the total fuel flow injected (in mg / stroke), and the gas fraction. burned in the intake manifold (%).

The NOx nitrogen oxide flow rate estimator 1 includes a map 2 of NOx nitrogen oxide flow rate inputted to the N rotation speed of the internal combustion engine and to the fuel flow rate Qe. The estimator 1 also comprises shape function correction means 3 as a function of the fraction Xgb of flue gas in the intake manifold. Alternatively (not shown), the estimator may be designed to require only the rotational speed of the engine (in rpm), the total fuel flow injected (in mg / stroke), the ratio of a box of the motor vehicle equipped with the engine (without unit, for example from 1 to 6 for a six-speed gearbox), and the fraction of gases burnt in the intake manifold (in%). It then comprises a map 2 of NOx nitrogen oxides flow which is connected in input to the rotation speed N of the engine, to the fuel flow Qe and to the transmission ratio BV which is engaged by the driver of the vehicle.

A multiplier 4 is connected to the outputs of the map 2 and the correction means 3 and outputs the estimation of the NOx nitrogen oxide flow rate of the internal combustion engine.

The cartography 2 is a function of the engine rotation speed, denoted N (in rpm) and the fuel flow rate noted Qe (in mg / stroke) and can be determined according to road tests, at rotational speed. constant motor, constant torque and no EGR.

Indeed, after the analysis of numerous stabilized and transient test bases, for a given engine and a given control calibration, the inventors have realized that the production of NOx nitrogen oxides can be considered almost constant. for a given speed and fuel flow.

In a first step, a NOx nitrogen oxide flow rate value is determined as a function of the engine rotational speed and the fuel flow rate through the oxides flow rate map 2. NOx nitrogen. The value of NOx nitrogen oxide flow rate can, in addition, be determined according to a transmission ratio BV engaged on the vehicle.

In a second step, the NOx nitrogen oxide flow rate values obtained from the mapping are corrected when the EGR is activated as a function of a shape function depending on the fraction of gases burned in the collector. admission. At the points of the map, for which exhaust gas recirculation (EGR for Exhaust Gas Recirculation) is active, NOx nitrogen oxide flow values must be reduced. For this, we use a function of function function of the fraction Xgb of burnt gases in the intake manifold. It is noted that the fraction of gases burned in the intake manifold is closely related to the rate of EGR.

Such a shape function can be determined as follows. From NOx nitrogen oxide flow rate measurements at constant engine rotation speed, at constant torque, and with EGR, the flow of NOx nitrogen oxides is again determined for the points corresponding to the points for which the flow rate of nitrogen oxides NOx has been previously determined at constant engine rotation speed, at constant torque, but without EGR. The ratio between NOx NOx flow without EGR and NOx NOx flow with EGR as a function of the burned gas fraction in the intake manifold is recorded as a scatter plot. The average curve of this scatter plot is the shape function. It is noted that the fraction of burned gases can be reconstructed simply from the rate of EGR and wealth.

The main difficulty in this estimation of NOx nitrogen oxide flow is to know precisely the fraction of gases burned in the intake manifold Xgb.

However, there is a simple relationship between the fraction of gases burned in the intake manifold Xgb and the composition of the gases LCG in the intake manifold (% of fresh gas):

Xgb = 1 - Fcol (Eq.1) The estimation of the composition of the Fccc gases in the intake manifold can then be determined by any direct or indirect means. However, reference may be made to the patent application FR2973441 to find an example of such an estimate of the composition of the LCG gases in the intake manifold.

It is also possible to use the system for determining the mass fraction of fresh gas in the intake manifold described below in relation to FIGS. 2 and 3.

In FIG. 2, a power train can be seen on which an estimate of the fresh air fraction Fcol in the intake manifold gases can be made. One can thus see an internal combustion engine 5 provided with an intake manifold 6 and an exhaust manifold 7 connected together by a high pressure partial exhaust gas recirculation pipe 8, provided with a valve, referred to as high pressure, referenced 9.

The exhaust manifold 7 is also connected to the turbine 10a of a turbocharger 10. The turbine 10a is also connected to the exhaust pipe 11, via a particulate filter 12. Alternatively, a catalyst Nitrogen oxide trap type (Noxtrap) or selective nitrogen oxides reduction catalyst (SCR) can also be connected at the outlet or inlet of the particulate filter.

The exhaust pipe 11 is connected to a partial recirculation pipe of the low-pressure exhaust gas 13, provided with a so-called low-pressure valve, referenced 14. The partial exhaust gas recirculation pipe at low pressure pressure 13 is connected at its other end to the compressor 10b of the turbocharger, via the fresh air intake duct 15.

The mass fraction F of fresh gas is variable in each of the volumes of the intake and exhaust line as a function of the richness in the combustion chamber, the opening of the HP EGR valve 9 and the opening of the the BP EGR valve 14. The mass fractions of fresh gas are then defined in each of the high-pressure and low-pressure volumes. The state variables used in the observer are as follows: F _avt : Mass fraction of fresh gas in the exhaust manifold 7

F _avc : Mass fraction of fresh gas upstream of the compressor

10b

Fcoi: Mass fraction of fresh gas in the intake manifold 6

Only the mass fraction of fresh gas in the exhaust manifold 7 can be determined directly, either by measurement thanks to a wealth probe, by an estimator depending, in particular, on the various operating parameters of the engine.

The evolution of these three mass fractions can be approximated by the following equations.

R- T ^A avt

F _avt = (Q _B F _∞1 -Q _f PCO- (Q _B + Q _f ) F

P ^R avt ^{• T} V ^T avt _a J

R • T ave

F _avc = ^ Q _AIR - (1-F _AVC ) + Q _LEGR - (F _avt -F _avc )) (Eq.2)

P pr aavvee ^■ · VV aavvee

RR- · T T

F∞i ^- (Qavc ^■ ^■ _stroke F (t ^- t _d) + Q _Hegr _avt · F - Q _B · F _col)

and

QB ⁼ Qair + QL QH ⁼ Qavc + QH

With:

R = Perfect gas constant (J / K.kg)

T = Temperature (K)

P = Pressure (Pa)

V = Volume (m ³ )

Q = Flow (Kg / s)

F = Mass fraction of fresh gas

avt = Exhaust Manifold

ave = compressor upstream

collar = intake manifold

Q _a i _r = Air flow through the air filter (Kg / s)

QLEGR = EGR Flow BP (Kg / s) Q _HE G _R = Flow rate E GR HP (Kg / s)

Q _avc = Flow upstream of the compressor.

Q _f = Fuel flow (Kg / s)

PCO = Combuivorous potency

ta = Delay time between compressor upstream and intake manifold.

These three equations are obtained by carrying out a mass conservation report relating to each of the mass fractions F _avt , F _co i, F _avc . A mass balance involves subtracting the mass outflows from the mass inflow. There is mass conservation if the mass quantity between the input and the output of the system remains constant despite the incoming and outgoing flows.

It should be noted that in the present case, the inflow and outflows do not represent the total flows, but only the flow rates of the gas corresponding to the mass fraction F considered. The gases coming from the upstream side of the compressor 10b, the intake manifold 6 and the exhaust manifold 7 are thus considered.

The mass fraction equation F _co relative to the intake manifold 6 involves a mass flow from the compressor 10b corresponding to the sum of the flow Q _a i _r in the fresh air duct and the flow Q _{LE GR} low pressure EGR 1 3. This flow must follow a relatively long path, which causes your delay that is to be determined.

An approximate value of this delay time ta can be determined by determining the transport time required to fill the volume between the upstream of the compressor 10b and the intake manifold.

6 with a given flow rate.

We define the mass m of this volume as follows:

P · V

m = (Eq.4)

R T

We define the transport time, equivalent to the delay time ta, in this volume as follows: m PV

(Eq

Qm R T Qm

With

Qm = mass flow

However, the introduction of this delay in an on-board calculation is extremely complex. Indeed, the delay time ta is not constant which has the effect of introducing a strong non-linearity in the calculations. Nonlinear calculations with infinite dimension being very heavy, they are therefore very long to solve.

To avoid such a situation, we will try to simplify the expression to obtain the delay time.

The delay of F _avc in the expression of F _co i is modeled by applying a discretization on the infinite dimension system. For this, we introduce a new equation Fi in the system of equation (Eq.2). The dynamic equation of Fi makes it possible to discretely model the space comprised in the pipes between the compressor 10b and the intake manifold 6. In other words, the space between the compressor 10b and the collector of admission 6 is divided into successive volumes Vi, in each of which a mass fraction Fi is determined. The index i thus varies from 1 to N, where N is the total number of discretizations.

The delay time ta, in the expression of the mass fraction F _co i relating to the intake manifold 6, is then replaced by the following approximate expression:

F _avc - (tt _d ) = F _N (Eq.6)

This equation amounts to asking that the value of the molar fraction of fresh gas upstream of the compressor is equal to the mass fraction Fi discretized for the last value i = N.

It is then postulated that in all these control volumes Vi, the mass fraction Fi is different, but that the pressures Pi and temperatures Ti are identical and equal. respectively at the pressure P _co i and the temperature T _co i in the intake manifold.

It also assumes that the Q _AVC upstream flow of the compressor 10b remains equal to the sum of the _flow rate Q _r i in the fresh air intake duct and the flow rate Qi _egr of the low-pressure EGR 13.

This involves rewriting the mass fraction equations (Eq.2) as follows:

F _avt = (Q _B F _co1 -Q _f PCO- (Q _B + Q _f ) -Fj

R · T

Favc ⁼ ~~ (Q air ^' (l ^_ ^ avc) ⁺ Q ^' (^ avt ^~~ ^ avc))

(Eq.7) -F)

R · T

F _C ol ⁼ (Qavc ^' FN + QH ^' F _avt ^- QB ^' F _∞ l)

With:

Fi: discrete mass fraction

V, = V _sura i / N

V _S urai: volume between the upstream compressor and the intake manifold

However, the equation system (Eq.7) remains nonlinear and therefore difficult to integrate into an onboard computer. To circumvent this difficulty, the method of linear systems with variant parameters (acronym "LPV") is applied. This method makes it possible to obtain an approximate solution of a nonlinear equation by solving a system of linear equations.

For this, we rewrite the system of equation (Eq.7) F _avt = P ₁ F _col -Q _f .Ρ0Ο- (ρ _{1 +} ρ ₂ ) Ε

P ( ^{"AVC F)} + P _^3" ^(F avt ^"F AVC) (Eq.8)

^F collar = P ₅ ^"F N + P ₆ - ^F avt ^" (P ₅ + P ₆ ) - ^F collar

With:

R ^■ T ^ avt

P ₂ = ^^ Q _f

PPaavvtt ^• V * ^v aavvtt

R ^■ T

₌ ^RT aavvcc Q

F3 _{T r} VLEGR

r P aavvcc ^■ V aavvcc (Eq.9)

₌ ^R R ^T Taavvcc Q

R · T ¹ collar

Po ^- ₇ HEGR

The values pi to p ₆ are also parameters varying from the linear model.

Taking into account the rewriting of the equations describing the system to equation Eq.8 and equations Eq.9, the system can be modeled by equations taking the following form:

X = A (p) .X ₊ W (p)

Y = C X

With:

p: the parameters varying from the linear model,

X: vector of state variables,

X: time derivative vector of the state variables,

C: the output matrix, A (p): first system state matrix, and

W (p): second system state matrix.

The state variables are included in the given vector X: X = [F _avt _stroke F F _t F2 ... F _N FF _col] ^T (Eq 1. 1)

The state variables correspond to the mass fractions in each of the control volumes defined between the upstream of the compressor 10b and the intake manifold 6.

The state variable Fi defined in the equations Eq. 7 and Eq. 8, here takes discrete values ranging from i = 1 to i = N.

The output matrix "C" is defined as follows: C = [l 0 · · · 0] This matrix C allows to reconstruct the mass fraction F _avt object of the discretization from different mass fractions (Eq. 12) Fi which are also, it is recalled, state variables included in the vector X. The equation Y = CX thus makes it possible to converge the system, F _avt being the only value of the system whose value is available.

There is a direct relation between F _avt and the state vector X. Indeed, the first scalar of the vector of the variables of state X is none other than the mass fraction F _avt .

It is thus possible to reformulate the equation Eq. 10 in the form of an observer in the following form:

Χ = Α (ρ) · Χ + W (P) + L (Y - Y) (Eq.13)

With

L: a gain matrix

X = the time derivative of the observer values for the vector X

X = the values of the observer for the vector X

Y = the value of the observer for the control value To ensure a good level of robustness of the observer, the gain matrix "L" is chosen so that the following conditions are met on the polytope defined by the extreme values, both minimum and maximum parameters pi to p ₆ .

A _; ^T P + PA _; + C ^T Y + Y ^T C + I <0

X> W ^1/2 YP ¹ · Y ^T · w ^1/2

min {Tr (VP) + Tr (x)} (Eq. 14)

P> 0

L = YP ¹

What is equivalent to the following representation: A _; ^T P + PA _; + C ^T · Y + Y ^T · C + I <0

min {Tr (VP) + Tr (x)} (Eq.15)

P> 0

L = YP ¹

With this observer, it is possible to obtain a stable estimate of the mass fraction of fresh gas. Figure 3 illustrates a system for determining the mass fraction of fresh gas based on the above equations.

A subtractor 16 receives as input the value of the mass fraction F _avt fresh gas in the exhaust manifold 7, measured or estimated, and the observer's value for the same mass fraction F _avt by applying the equation eq. 10.

A means 17 for determining the gain of the observer receives as input the difference between the measured value and the value of the observer of the mass fraction F _avt . It determines the gain matrix L by application of equation Eq. Then the term L (Y-Yj of equation Eq.16.

A first calculation means 19 determines the first state matrix A (p) as a function of the equations Eq. 8 and Eq. 10, while a second calculating means 20 determines the second state matrix W (p) as a function also of the equations Eq.8 and Eq. 10.

For this purpose, the first calculation means 19 and the second calculation means 20 receive as input the measurement P _oc i of the pressure in the intake manifold, the measurement T _co i of the temperature in the intake manifold, the measurement of the intake air flow rate Q _a i _r , the flow measurement QHEGR of the HP EGR, the measurement F _avt of the richness in the exhaust manifold, the estimation of the QLEGR flow rate of the BP EGR, the estimate P _with the pressure upstream of the compressor, the estimate _Tcc of the temperature upstream of the compressor and the setpoint Q _f of the fuel flow rate imposed by the control of the engine.

The measurement of the intake air flow Q _a i _r can be carried out for example by a flowmeter disposed at the level of the air filter.

The QHEGR flow measurement of the HP EGR can be performed by differential pressure across the HP EGR valve.

Based on these values, the first calculation means 19 and the second calculation means 20 determine the values pi to p ₆ by applying the equations Eq.9.

The first calculation means 19 and the second calculation means

Then determine respectively the matrices W (p) and A (p) by applying the equations Eq.8 and Eq. 10.

An adder 18 receives the term L (Y-Y) from the determination means 17, the first state matrix A (p) from the first calculation means 19 and the second state matrix W (p) from the second calculation means 20.

An integration means 21 receives from the summator 18 a vector

X temporal derivatives of observed variables from equation Eq. 13.

A third calculation means 22 receives integration means

21 a vector of the observed variables X and outputs the value of the mass fraction F _co i at the intake manifold by applying equations Eq. 1 1 and Eq. 13.

The vector of the observed variables X is also transmitted to the second calculation means 20 and to a fourth calculation means 23. The fourth calculation means 23 transmits the value of the observer of the mass fraction F _avt to the subtractor 16 by applying the equation Eq. 10.

We can thus estimate the mass fraction F _co i deduced from the vector of the observed variables X, from the difference between the measurement and the estimate of the mass fraction F _avt = Y.

The choice of the number of discretizations "N" depends on:

the available calculation time, because for each increase of "N", a new state equation is introduced,

- the desired precision for estimating the mass fraction F coi under transient conditions,

- the robustness of the estimate, especially on the very strong transient regimes with the risk of oscillations for sudden input variations (slot type).

For the application described, the value "N = 3" seems to be a good compromise for the desired services with the constraints imposed.

It is thus possible to correctly estimate the mass fraction of fresh (or burnt) gases in the intake manifold of an engine equipped with a low pressure EGR. In particular, it is possible to take into account the delay of the transport phenomena in the intake line. This taking into account of the delay in the observer makes it possible to guarantee a good level of prediction. In addition, the passage through the LPV approach and the definition of stability on the polytope defined by it guarantees the robustness of the observer. This makes it possible to have an estimate even in critical rolling cases compared to known solutions.

Thus, during the second step, before the application of the shape function to correct NOx nitrogen oxide flow rates, the mass fraction of fresh gas Fco 1 is determined so that determine the value of the fraction of burnt gases in the intake manifold Xgb.

This implies that the system for determining the mass fraction of fresh gas is connected at the input of a subtractor able to determine the fraction of gases burned in the intake manifold Xgb from the mass fraction of fresh gas Fcol by application of the equation Eq. 1. The output of the subtracter is then connected to the input of the correction means 3. It will be noted that the system for determining the fresh gas mass fraction and the subtractor are not illustrated in FIG. 1.

The system for determining the mass fraction of fresh gas is previously calibrated in order to obtain coherent values, the calibration being carried out by supplying the geometrical quantities of the engine air chain under study.

The flow rate of the nitrogen oxides NOx is thus estimated from the measurement of the exhaust richness, the engine control variables such as the fuel flow and the ignition advance and the estimation of the composition of the gases in the intake manifold.

According to another embodiment illustrated in FIG. 4, the NOx nitrogen oxide flow rate estimator 1 comprises an algebraic relationship estimation means 24 replacing the mapping 2 and the form function correction means 3. The estimation means 24 makes it possible to take into account more phenomena in order to increase the level of prediction. The algebraic relation is of the following form:

NOx = min [F ₁ , F ₂ ] -Qe (Eq.16)

F _j = (aR + b)

F ₂ = (a + p-Qe + x-Qe ² + δ-Qe ³ ) - ^ - exp (v-Xgb)) - (ro + 9-ADV)

With:

a, b, α, β, χ, δ, μ, ν, θ, ω: parameters of the estimator

Qe: fuel flow

R: wealth

Xgb: fraction of burnt gases in the intake manifold ADV: phase of the main injection The inputs of the estimation means 24 by algebraic relation are the richness R, the fuel flow rate Qe (in mg / stroke), the fraction Xgb of flue gases in the intake manifold, and the ADV phase of the main injection. .

As for the first embodiment, the fraction Xgb of flue gas is determined from the equation Eq. 1 from the Fcol value from the system for determining the fresh gas mass fraction.

According to the algebraic relation (Eq.16), we can see that the rate of oxides of nitrogen corresponds to the minimum value among the values taken by two functions, denoted Fi and F ₂ . The first function Fi depends solely on the fuel flow Qe and the richness R. The second function F ₂ depends on the fuel flow Qe, the phasing of the main injection ADV and the fraction of burnt gas Xgb.

During the calibration of the estimation means 24, the parameters of this algebraic relationship are determined by comparing the flow rate of the NOx nitrogen oxides measured at the flow rate of the modeled nitrogen oxides NOx, for common values of richness R, of flow rate. Qe fuel, Xgb fraction of flue gas in the intake manifold, and ADV phase of the main injection. By way of example, the values taken by the coefficients of this expression are generally the following:

a = -0.0865, b = 0.1023

α = -0.020, β = 0.000496, χ = -1.016ε-6, δ = 0.7983ε-7

μ = 3.7121, ν = -0.1121

ω = 1.966, θ = 0.0301

The result level of the two embodiments of the estimators is substantially equivalent.

Claims

1. System for estimating the flow of nitrogen oxides in the exhaust gases of an internal combustion engine for a motor vehicle equipped with at least one partial exhaust gas recirculation circuit and at least one an exhaust gas reprocessing system, the flow of nitrogen oxides being estimated upstream of exhaust gas reprocessing systems, characterized by the fact that it comprises

a system for determining the mass fraction of fresh gas,

a subtractor capable of determining the fraction of gases burned in the intake manifold from the mass fraction of fresh gas, and

an estimator (1) of the flow rate of nitrogen oxides as a function of the rotational speed of the internal combustion engine and the fraction of burnt gases in the intake manifold.

The system of claim 1, wherein the estimator

(1) the flow of oxides comprises a map (2) of the flow of nitrogen oxides connected at input to the rotation speed of the internal combustion engine and the fuel flow, and a means of correction (3 ) by function of form as a function of the fraction of gases burned in the intake manifold, and a multiplier (4) connected as input to the outputs of the map (2) and the correction means (3), able to transmit in output estimation of the flow of nitrogen oxides of the internal combustion engine.

The system of claim 2, wherein the mapping

(2) the flow of nitrogen oxides is further connected in input to a gear ratio (BV) engaged on the vehicle.

4. The system of claim 1, wherein the estimator (1) of the flow of nitrogen oxides comprises a means of estimation (24) by algebraic relationship depending on the richness, the fuel flow, the fraction of burnt gas in the intake manifold, and the phase of the main injection.

5. A method for estimating the flow of nitrogen oxides in the exhaust gas of an internal combustion engine for a motor vehicle equipped with at least one partial exhaust gas recirculation circuit and at least one an exhaust gas reprocessing system, the flow of nitrogen oxides being estimated upstream of exhaust gas reprocessing systems, characterized in that it comprises the following steps:

6. Method according to claim 5, wherein the flow rate of nitrogen oxides is determined according to a mapping function of the engine rotation speed and the fuel flow, then

the nitrogen oxide flow rate values are corrected according to a shape function depending on the fraction of gases burned in the intake manifold.

7. The method of claim 6, wherein the mapping is further function of a gear ratio engaged on the vehicle.

8. Process according to claim 5, in which the rate of nitrogen oxides is determined as a function of an algebraic relationship and of the fuel flow, the richness, the phasing of the main injection and the fraction of gas. burned.

9. Process according to any one of claims 5 to 8, wherein, based on measurements of nitrogen oxide flow rate constant engine rotation, constant torque, and with partial recirculation of exhaust gas. the flow rate of nitrogen oxides is determined for operating points corresponding to the operating points for which the flow of nitrogen oxides has previously been determined without partial recirculation of the exhaust gas, the ratio between the flow rate of nitrogen oxides without partial recirculation of the exhaust gases and the flow rate of nitrogen oxides with partial recirculation of the exhaust gases is memorized for a set of values of the fraction of gases burned in the intake manifold, and

the function of form is determined as being the average function of all these points.