EP0961875A1

EP0961875A1 - Method for controlling the purge of a catalyst container treating an internal combustion engine exhaust gases

Info

Publication number: EP0961875A1
Application number: EP98959959A
Authority: EP
Inventors: Stéphan BEURTHEY; Daniel Cozler; Jean-Claude Mollet
Original assignee: Renault SAS
Current assignee: Renault SAS
Priority date: 1997-12-12
Filing date: 1998-12-11
Publication date: 1999-12-08
Anticipated expiration: 2018-12-11
Also published as: DE69811863T2; EP0961875B1; WO1999031368A1; DE69811863D1; FR2772428A1; ES2190123T3; FR2772428B1

Abstract

The invention concerns a catalyst container (6) comprising means absorbing nitrogen oxides contained in said gases and a method consisting in starting a purge by increasing the richness (R) of the engine (1) air/fuel mixture supply, on the basis of a richness corresponding to a lean or stoichiometric mixture; evaluating, using a table based on the container (6) temperature (T>cat<) and rate of filling (NS/NSC) with nitrogen oxides, the efficacy (EFFNOX; EFFSTOCK) of the container in absorbing nitrogen oxides when the engine (1) operates with lean mixture, and in starting a purge of the container (6) when said efficacy (EFFNOX; EFFSTOCK) falls below a predetermined value (Eff_min), based on the container (6) temperature (T>cat<). The invention is applicable to an internal engine powering a motor vehicle.

Description

Method of controlling the purging of a catalytic converter for treating the exhaust gases of an internal combustion engine

The present invention relates to a method for controlling the purging of a catalytic converter for treating the exhaust gases of an internal combustion engine comprising means for adsorption of nitrogen oxides contained in these gases, type according to which such a purge is triggered by an increase in the richness of the air / fuel mixture supplying the engine, from a richness corresponding to a lean or stoichiometric mixture. To reduce the pollution of the atmosphere by the exhaust gases of internal combustion engines, in particular those which ensure the propulsion of motor vehicles, the exhaust lines of these gases are commonly fitted with a catalytic converter having the function transform the harmful chemical species contained in these gases into less harmful, even harmless, chemical species.

This is how a catalytic converter known as "three-way" or "trifunctional" is known, because it performs a triple function of oxidation of unburnt hydrocarbons (HC), of oxidation of the dangerous CO to C0 ₂ , and reduction of nitrogen oxides to nitrogen gas. The concern to further reduce atmospheric pollution has caused interest in an internal combustion engine capable of operating essentially with an air / fuel mixture called "lean", that is to say whose fuel richness is less than that of the stoichiometric mixture. When a conventional three-way catalytic converter is used to clean up the exhaust gases of such an engine operating in a lean mixture, there is however a clear loss of efficiency of the pot with regard to the reduction of nitrogen oxides (noted in the following NOx), which are particularly harmful.

To overcome this drawback, it is proposed in European Patent No. 560,961 to use a catalytic converter trifunctional modified in that it comprises means for adsorption of NOx present in the exhaust gases, when the air / fuel mixture supplied to the engine is lean, which avoids then evacuating them into the atmosphere. To prevent this adsorption from causing, over time, a saturation of the adsorption capacity of the pot with NOx, and therefore a loss of efficiency of the pot from this point of view, the aforementioned European patent proposes to control the richness of the air / fuel mixture supplying the engine so as to periodically switch the richness of this mixture from a value corresponding to a lean mixture to a value corresponding to a stoichiometric or superstoichiometric mixture, this switching causing the desorption of the NOx adsorbed in the pot then their reduction by HC and carbon monoxide CO then in the pot due to the increase in the richness of the mixture, these last two chemical species must themselves be oxidized into less polluting species, in accordance with two of the three functions classics of the catalytic converter.

This switching of the richness of the air / fuel mixture from a value corresponding to a lean mixture to a value corresponding to a rich mixture, for the purpose of depolluting the exhaust gases, leads to an overconsumption of fuel depending on the means used. to assess the quantities of gas adsorbed (or "stored") and desorbed (or "destocked") by the pot as well as the means used to trigger the purge of the pot and regulate the duration of this purge. In this respect, a device of the type mentioned above is known from European patent application No. 598 917 which develops a strategy according to which it is considered that below a certain stored quantity of NOx, the NOx arriving in the pot are completely stored. Above this quantity, only a fraction of the NOx arriving in the pot is stored.

The decision to purge the pot is made when the NOx pot filling rate is exceeded. However, the NOx adsorption capacity of the pot is a function of the temperature of the pot, not taken into account in this strategy. In addition, the duration of a purge depends only on the temperature of the pot. It therefore does not take into account another essential variable, namely the flow rate of the engine in reducing materials (HC, CO).

Also known from European patent application No. 733 786 is a device comprising, downstream of the pot, an oxygen sensor. During a purge, when this probe indicates that the richness of the air / fuel mixture supplying the engine, evaluated from the composition of the exhaust gases, is greater than 1, the purge of the pot is considered to be complete and is arrested. The purge time is then a function of the quantity of NOx destocked, which is a function of the engine speed, the opening time of fuel injectors in the engine, the richness of the air / fuel mixture and a function parameter. the deterioration of the pot, taken into account by the signal delivered by the oxygen sensor. This device has the drawback of requiring the presence of an oxygen sensor downstream of the catalytic converter.

The object of the present invention is to provide a method for controlling the purging of a catalytic converter for treating the exhaust gases of an internal combustion engine comprising means for adsorption of nitrogen oxides contained in these gases. , a process which does not have the drawbacks mentioned above of the processes for purging catalytic converters of this type known in the prior art, and which in particular makes it possible to further reduce the overconsumption due to the purging of the NOx pot by closer monitoring storage and destocking of NOx in the catalytic converter and by a more precise measurement of the purge times of the pot.

The present invention also aims to provide such a process which does not require the presence of an oxygen sensor at the outlet of the catalytic converter.

These objects of the invention are achieved, as well as others which will appear on reading the description which follows, with a method of controlling the purging of a catalytic converter for treating the exhaust gases of an engine with internal combustion, said pot comprising means for adsorption of nitrogen oxides contained in these gases, according to which such a purge is triggered by an increase in the richness of an air / fuel mixture supplying the engine, starting of a richness corresponding to a lean or stoichiometric mixture, this process being remarkable in that a) one evaluates, using a table depending on the temperature of the pot and its filling rate with nitrogen oxides, the pot's efficiency in terms of adsorption and reduction of nitrogen oxides, and b) a purge of the pot is triggered when said efficiency falls below a predetermined value, depending on the temperature of the pot.

According to the invention also, said rate of filling of the pot with nitrogen oxides is evaluated on the basis of an evaluation of the current quantity of oxides stored during an operating phase of the engine in lean mixture, said rate is taken from this evaluation and a table giving the maximum quantity of nitrogen oxides adsorbable by the pot, according to the temperature of the pot and the current concentration of nitrogen oxides in the engine exhaust gases.

For the evaluation of the current quantity of nitrogen oxides stored, the EFFDENOX ratio between the flow of nitrogen oxides treated by the jar and the flow of nitrogen oxides leaving the engine is taken into account. report EFFDENOX being taken from a table depending on the temperature of the pot.

During a purge, the quantities of nitrogen oxides destocked successively by the pot are continuously evaluated by the quantities of fuel successively injected into the engine and the purging is stopped when the efficiency of the pot rises above said value of predetermined threshold due to said successive destocking.

According to the invention also, the overall instantaneous efficiency of the pot is measured by the EFFNOX ratio of the flow of nitrogen oxides at the outlet of an engine exhaust line, to the flow of said oxides at the outlet of the engine.

As a variant, said efficiency is the instantaneous EFFSTOCK storage efficiency such that:

EFFNOx-EFFDENOX EFFSTOCK =

1-EFFDENOX

Other characteristics and advantages of the present invention will appear on reading the description which follows and on examining the appended drawing in which: - Figure 1 is a diagram of an internal combustion engine equipped with means allowing executing the purge control method according to the invention, FIG. 2 is a graph of the overall instantaneous efficiency EFFNOX as a function of the filling rate and the temperature of the pot,

FIG. 3 is a graph of the maximum storage capacity NSC of the pot,

- Figure 4 is a graph of EFFDENOX efficiency as a function of the temperature of the pot, - Figure 5 is a graph of the dNS / dM ratio of the amount dNS of NOx destocked by an amount dM of reducing materials present in the pot , depending on the filling rate of the pot, at a pot temperature of 350 ° C. - Figure 6 is the flow diagram of an expert system implementing the control method according to the present invention, and Figures 7 to 12 are groups of timing diagrams illustrating various phases of the control method according to the present invention.

There is shown in Figure 1 of the accompanying drawing an internal combustion engine 1 controlled by a digital electronic computer 2. Conventionally, the computer 2 receives and processes signals from various sensors, in particular from a sensor 3 of the pressure P (or flow) of the air admitted into the engine 1, to control in particular the opening time ti of fuel injectors 7 in the engine 1. The engine exhaust gases pass through an exhaust line 4ι, 4 ₂ , 4 ₃ which successively passes through, and in this order, a catalytic pot of possible priming 5 and trifunctional catalytic pot 6 of the type of those mentioned in the aforementioned patents, namely a pot comprising means for adsorption of nitrogen oxides NOx when the engine operates in a lean mixture, these oxides being desorbed, or "destocked", when the engine is supplied with an air / fuel mixture of richness greater than or equal to 1 suitable for leaving in the engine exhaust gas reducing chemical species (HC, CO) capable of reducing the desorbed NOx, unburnt hydrocarbons HC and CO being themselves then oxidized into less polluting or non-polluting chemical species (C0 ₂ , H ₂ 0, etc ...). A temperature sensor 8 supplies the temperature T _cat of the pot 6 to the computer 2.

This digital computer 2 is duly programmed to execute, in addition to the tasks which are conventionally entrusted to it, the method according to the invention for controlling the purging of the catalytic converter 6, and for treating NOx desorbed during the purges, which will be described below.

This process makes use of models for the evaluation of the quantities of nitrogen oxides which are successively stored by the pot, destocked and reduced by this pot.

It also makes use of an expert system to control the triggering and the duration of the purges according to the evaluations drawn from the models. These models and this expert system implemented successively using software means incorporated in the digital computer 2 are described successively.

For the modeling of the operation of the catalytic converter 6, the following operations will be distinguished: - operation in storage of nitrogen oxides, hereinafter called "NOx storage", operation which is observed when the engine is supplied with lean mixture ,

- operation in destocking of nitrogen oxides, or "destocking of NOx", which is observed when the engine is supplied with a rich mixture, the destocked NOx being then reduced thanks to the excess of hydrocarbons (HC) and monoxide carbon (CO) that is found in the engine exhaust due to the incomplete combustion of the air / fuel mixture, and

- operation at stoichiometry, where it is assumed that the pot 6 does not store or destock NOx. In this regard, it should be understood that a "stoichiometric" air / fuel mixture supplying the engine can however produce exhaust gases containing unburnt hydrocarbons and CO owing to the incomplete combustion of the input stoichiometric mixture. These reducing compounds can then ensure the destocking and reduction of NOx, just like a "rich" mixture. This is why we will mainly distinguish in the following the mixtures "poor" air / fuel with a richness less than 1 and mixtures with a richness greater than or equal to 1. NOx storage operating model

The pot 6 then adsorbs part of the NOx formed in the engine, reduces another part of it thanks to any unburnt hydrocarbons and carbon monoxide and releases into the exhaust line 4 a third part of these NOx, corresponding to oxides of which are neither stored nor reduced. According to the present invention, the overall instantaneous efficiency EFFNOX, the instantaneous treatment efficiency EFFDENOX and the instantaneous storage efficiency EFFSTOCK of the pot are defined as follows: NOx flow rate at the outlet of the exhaust line

EFFNOX = —— -; NOx flow at motor output 1

_ reduced NOx flow in the pot 6 EFFDENOX NOx flow at the motor outlet 1

EFFNOX - EFFDENOX

EFFSTOCK =

1- EFFDENOX

We also define the filling rate NS / NSC of pot 6: quantity of NOX adsorbed (NS)

NS / NSC maximum adsorbable quantity (NSC) Measurements carried out on the engine on the bench made it possible to detect variations in EFFNOX, EFFDENOX and NS according to various parameters. EFFNOX is a function of the temperature of the pot 6, the filling rate NS / NSC as well as, to a lesser extent, the flow rate of the engine in NOx. The graphs in Figure 2 show variations in EFFNOX depending on the temperature and the filling rate.

FIG. 3 shows the shape of the variations in NSC as a function of the temperature of the pot, for two concentration values [NOxJi and [NOx]; _. nitrogen oxides in the gases leaving the exhaust line.

FIG. 4 shows the variations of EFFDENOX as a function of the temperature of the pot.

It is understood that the various graphs above can be stored in the computer 6 in the form of tables or "maps", with one or more entries, the reading of these tables providing values of the quantities and variables taken into account in the operating model, according to the invention, of the pot 6 for NOx storage which is described below.

This operation of the pot 6 occurs when the air / fuel mixture supplying the engine is poor, that is to say sub-stoichiometric. If NS__ is the quantity of nitrogen oxide adsorbed at the instant of sampling î (corresponding for example, as is conventional, to the change to top dead center TDC _3. Of a cylinder of engine 1), we have : NSj _. = NSχ- ₁ + DNSSTOCK (1)

DNSSTOCK being the quantity of NOx stored between the TDCs of order (1 - 1) and 1.

According to the invention, DNSSTOCK is evaluated by:

DNSSTOCK = [NOx]. (1 - EFFDENOx). EFFSTOCK, i.e. DNSSTOCK = [NOx]. (EFFNOx - EFFDENOX) (2) where [NOx] is the NOx concentration, ie the ratio of the mass of NOx produced by the engine to the mass of air admitted to each combustion of this engine. A map (not shown) makes it possible to evaluate this mass of NOx at each combustion, as a function of the speed N and of the engine load, with application of a correction as a function of the temperature of the engine cooling water. The air mass is evaluated by the computer 2, so conventional, from a measurement of the air flow or pressure at the intake manifold of engine 1.

The maps illustrated in Figures 2 and 4 provide the values of EFFNOX and EFFDENOX to be taken into account in the calculations of DNSSTOCK. The formulas (1) and (2) above make it possible to know at any time the quantity NSi_ι of NOx stored in the pot 6. The cartography of FIG. 3 gives NSC. We therefore know the NS / NSC ratio and the temperature T _cat of the pot, which makes it possible to derive EFFNOX from the cartography of FIG. 2. Operating model for destocking and reduction of NOx

According to the invention, this destocking and this reduction occur when the richness of the air / fuel mixture supplying the engine is greater than or equal to 1. Measurements on the bench made it possible to establish the graph of FIG. 5, which represents the variations in the dNS / dM ratio of the quantity of NOx destocked by a mass dM of reducing agents (H ₂ , CO, unburnt hydrocarbons) as a function of the filling rate NS / NSC of pot 6, for a temperature of this pot equal to 350 ° C. The graph in FIG. 5, and other similar ones corresponding to other values of the temperature, are stored in the computer 2 in the form of a map, like the graphs in FIGS. 2 to 4. According to the invention, the destocking occurs when the richness of the air / fuel mixture exceeds a RICPURGE threshold value, the quantity NSi of NOx remaining stored in the pot at TDC of order i being decremented by a quantity DNSPURGE corresponding to the quantity of NOx destocked between the TDCs (i - 1) and i, that is:

NSi ≈NSi-i - DNSPURGE, with: DNSPURGE + QNOXTRAPi. dNS / dM, where: dNS / dM is known by the cartography corresponding to the graphs of FIG. 5, and QNOXTRAPi is the mass of reducers received between two TDCs of order (i - 1) and i by the pot 6.

In this regard, it should be noted that, when the exhaust line comprises, as shown in FIG. 1, a priming catalytic converter 5 in front of the pot 6, the reducers which leave the engine 1 are subjected to oxidation in the pot 5 by the oxygen stored therein. It is only the excess quantity, not oxidized by this oxygen stored in the pot, which can be used in pot 6 to destock or reduce the stored NOx. QNOXTRAP is therefore equal to the mass of QESSi reducers delivered to the engine exhaust between two TDCs of order (i-1) and i when the combustion of the air / fuel mixture is incomplete and that the priming catalytic converter 5 is oxygen vacuum, either:

^QESSi .Cbitrate (3) or :

- RICONSi: setpoint richness during the purge,

- RICPURGE: richness of the air / fuel mixture below which there is no longer enough unburnt HC and CO in the exhaust gases to effectively purge the pot 6,

- DINJV: opening time of a fuel injector 6 in the engine 1, - C flow rate: scaling coefficient.

During a RICONSi> RICPURGE rich purge, we then have:

QNOXTRAPi = QESSi

calculated by formula (3) above if, in addition, there is then no injection cut off resulting from a request for deceleration of the vehicle by the driver and if the cumulation MESSi at PMHi of the quantities of reducers debited in pot 6 from the purge flow is such that:

MESSi = MESSi-i + QESSi> MESSLO

MEΞSLO being a threshold value of the quantity of reducing agents necessary to purge the priming catalyst 5 from the oxygen stored in this catalyst.

If the conditions defined above are not met, we have:

QNOXTRAPi = 0 MESSi = 0

In the foregoing, the influence of the various reducers (HC and CO) on the reduction of NOx is taken into account in a global manner through an equivalent reducer flow rate QNOXTRAP. We will now describe the structure of an expert system making it possible to make decisions as to the advisability of triggering a destocking of the NOx contained in the pot 6, that is to say a transition to so-called "PURGE" mode. The expert system also makes it possible to make a decision as to the advisability of stopping such a purge. Description of the expert system

According to a characteristic of the present invention, the taking of a destocking decision involves either the overall instantaneous efficiency EFFNOX, or the instantaneous storage efficiency EFFSTOCK.

Each of these two quantities can be taken into account either by its instantaneous value or by its current average value since the last purge.

The average values of these efficiencies, overall or storage, can be used to trigger a purge. On the other hand, they cannot be used to regulate the duration of the purge and therefore to stop it, their progress being too slow. With this restriction, the efficiencies mentioned below can be represented by one or the other of the instantaneous or average, storage or global efficiencies, defined above.

We will now describe the structure and operation of the expert system integrated into the purge control method according to the invention, in conjunction with FIG. 6 which presents the flow diagram of this expert system.

On it appear several binary digital information which have the following meaning: - lean_state = 1 indicates that the conditions necessary for the correct operation of engine 1 in lean mixture are met, for example: the engine speed and the pressure ratio d air to the manifold at atmospheric pressure are below respective predetermined thresholds, the temperature of the engine cooling water and the vehicle speed are greater than respective predetermined thresholds. Otherwise, lean_state = 0.

- auto_lean = 1 indicates that the conditions for a return to engine operation in a lean mixture are met, for example: the accelerator has left the "lifted foot" position for a time below a predetermined threshold where: the pressure variation at the engine manifold between two consecutive combustions is less than a predetermined threshold. Otherwise, auto_lean = 0.

- autretlean = 1 indicates that the purge has been triggered since a lean mixture operation and that a return to lean mixture operation of the engine is authorized, as soon as the conditions for such a return are met.

Thus, for example, the engine having left an operation in lean mixture to carry out a purge of the pot 6, returns to operation in mixture poor at the end of the purge if at this time autretlean = 1.

- état_nox = 1 indicates an authorization to switch to "purge" mode. ox_state = 0 indicates a purge stop authorization.

The flow diagram of FIG. 6 illustrates the conditions which govern the changes in the richness setpoint of the air / fuel mixture supplying the engine, as a function of the conditions which determine a switching of the operating mode of the engine between three modes: the mode "purge", the "stoichiometric mixture" mode and the "lean mixture" mode. It is the switches between any of these modes and any of the other two that are managed by the expert system. This is how it manages the triggering and stopping of the catalytic converter purge phases.

The purge operations of the catalytic converter 6 can be triggered while the engine is operating either in lean mixture or in stoichiometric mixture, it being understood that when the engine operates at a richness greater than 1, this operation is naturally accompanied by a purge of the pot.

Purging from the operation of the engine in poor mixture As indicated in the flowchart in Figure 6, the richness of the mixture is then fixed at a RICPURGE value taken from a "carlean" map depending on the pressure P at the manifold. engine intake and N speed of this engine. The models described above allow permanent monitoring of the evolution of the pot's efficiency, for example the instantaneous EFFNOX efficiency. When this falls below a predetermined threshold Eff_min_lean depending on the temperature of the pot, the system commands a purge of the pot (the nox state bit goes to 1). If as a result of an increase in the engine torque requested by the driver of the vehicle for example, the operating conditions for lean mixture are no longer met (lean_state = 0), the change to purge mode is authorized (nl_state = 1) when l pot efficiency falls below a threshold depending on the temperature of the pot.

According to the present invention, there is then an obligatory transition from the richness of the air / fuel mixture to a richness greater than or equal to 1 suitable for satisfying the torque demand, in order to purge the pot, which reduces the overconsumption of fuel due to bleeding. Purge, since the engine is running in stoichiometric mixture The engine then operates at a RICSTOC richness drawn from a "carstoc" map whose inputs are the pressure at the intake manifold and the engine speed.

During such an operation, there is in principle neither storage nor destocking of NOx except in the presence of a fraction of unburnt hydrocarbons due to incomplete combustion, as has been seen above. It does, however, predict what would happen if the composition of the air / fuel mixture were to return to that corresponding to a lean mixture. We then evaluate the instantaneous storage efficiency EFFNOX, for example, and we compare it to a low threshold value Eff_min. If this efficiency is lower than the threshold, we purge the pot by setting the status_nox bit to 1. Other conditions for switching to purge mode can be set. For example, if you do not want the purge to last too long, it will only be allowed if the engine load is above a threshold, in which case the mass d unburnt hydrocarbons present in the pot is important and capable of reducing the duration of the purge. Other conditions could be posed, with regard to the vehicle speed, the engine speed, the gearbox engaged ratio, the temperature of the engine cooling water, etc. etc. purge mode

When a purge is requested, the computer 2 reads a wealth of destocking or RICPURGE purge in a "cardenox" table, the inputs of which are the pressure P at the engine manifold and the speed N of the latter.

The computer then changes the richness setpoint of the air / fuel mixture, either from the RICSTOC value taken from the carstoc map if the engine was previously operating in stoichiometry mode, or from the RICLEAN value if the engine previously operated in lean mode, to a RICPURGE purge richness instruction taken from the "cardenox" table.

According to an advantageous embodiment of the present invention, it is then made so that the variation of the engine torque which results from the setting of this new wealth setpoint is entirely compensated by other means, so that the torque demand fixed by the driver does not vary, which would otherwise be unpleasantly felt by him and its possible passages, in the form of an uncontrolled variation of the speed of the vehicle. According to the present invention, this compensation is carried out either a) by an action on the angle of advance at ignition of the air / fuel mixture when the vehicle does not include means making it possible to automatically control the flow of incoming air in the mixture either b) by an action on such means, a "motorized" throttle controlling this air flow, for example.

We will now describe the development of a purge phase in the case of an engine not equipped with a like a "motorized" butterfly. We will successively distinguish three situations: 1) that of a purge triggered from a lean mixture operation, 2) that of a purge triggered by a loss of operating conditions in lean mixture and 3) that of a purge triggered by a insufficient storage efficiency, evaluated during engine operation at stoichiometry. 1) Purge triggered from an operation in poor mixture It is then an insufficient efficiency of the pot which triggers the purge (state_nox is set to 1). On the timing diagram of FIG. 7, it appears that the richness is then brought to RICPURGE (at time ti), the autretlean bit being set to 1. To keep the engine torque constant, the advance Av is simultaneously reduced to l 'ignition, so as to compensate for the increase in torque which results from the increase in the setpoint richness.

The effectiveness of pot 6 then rises gradually. As soon as this (at time t ₂ ) exceeds a threshold value Eff_max, depending on the temperature of the pot, the purge is stopped (nox_state is set to 0). The richness R returns to the RICLEAN set value stored in the carlean table, the autretlean bit is reset to 0 and the lowering of the torque following the lowering of the richness is compensated by a cancellation of the advance withdrawal ordered at the start purge.

If, during a purge as described above, the operating conditions for lean mixture disappear (autretlean bit set to zero), the advance Av is reduced to the value corresponding to the combustion of a stoichiometric mixture, with consumption minimum (see Figure 8, time t ₃ ). This return is progressive in order to render insensitive the increase in engine torque which results therefrom while the bleeding ends. When Eff efficiency has returned to value maximum threshold (instant t), the richness R is gradually reduced to stoichiometry, ie to the value RICSTOC.

2) Purge triggered by loss of operating conditions in poor mixture

The timing diagrams in Figure 9 illustrate this situation, which appears at time ts. The richness of the mixture then gradually changes to the RICPURGE value and the ignition advance is gradually decreased conco itingly. The purge thus triggered increases the efficiency of the pot to a threshold Eff_max reached at time tç. The purge is then stopped by a progressive command to return the richness R to stoichiometry (ie RICSTOC). Another possible evolution of the purge has been illustrated in FIG. 10. After triggering of the latter at time ts, as in FIG. 9, the conditions for returning the operation of the engine in lean mixture reappear at time t ₇ , while the purging is not completed (the auto_lean bits and lean_state are then 1, as shown in Figure 6). The wealth is gradually brought back to RICLEAN and the advance goes up to the corresponding value (instant ts).

3) Purge triggered from stoichiometric operation

FIG. 11 illustrates the development of such a purge, triggered by the transition from efficiency to a low threshold Eff_min at time tg. The richness then gradually rises to the RICPURGE value while the advance Av is maintained at the previous value suitable for operation at stoichiometry. In FIG. 11, the purging stops at the instant tio when the efficiency has returned to a high threshold Eff_max. Wealth is gradually brought back to RICSTOC. In FIG. 12, the stopping of the purge is triggered by the appearance, at time tu, of conditions for returning to an operation with lean mixture (auto_lean = 1 and état_lean = 1, see FIG. 6). The advance is then gradually increased to a value suitable for the combustion of a lean mixture.

It now appears that the invention makes it possible to achieve the set goals, namely to provide a process for controlling the NOx purge of a catalytic converter comprising means for adsorption / desorption of such oxides contained in the exhaust from an internal combustion engine, capable of organizing this purge so as to minimize the overconsumption of fuel that it implies. The savings thus obtained by implementing this process are further increased by the fact that this process does not require the presence of an oxygen sensor at the outlet of the catalytic converter. Furthermore, the method according to the invention compensates for the variations in the engine torque that could otherwise be observed during the purging of the pot, for the benefit of the comfort of driving the vehicle and, more generally, of the comfort of the passengers of the vehicle.

Of course, the invention is not limited to the embodiment described and shown which has been given only by way of example. This is how compensation for the variations in torque mentioned above could be obtained other than by acting on the ignition advance. As seen above, in vehicles equipped with means for automatically adjusting the amount of air entering the engine, controlled by the engine management computer, such as a so-called "motorized" butterfly, for example, compensation for excess torque due to the enrichment of the air / fuel mixture during purges, compared to the setpoint set by the driver, can be obtained by a reduction concomitant with the amount of air entering the engine. The general strategy for initiating and stopping the purges described above in conjunction with the timing diagrams of FIGS. 7 to 12 remains substantially the same.

The present invention obviously extends to a control method in which the variations in torque mentioned above are compensated by a combined action on the ignition advance and on the quantity of air entering the engine, when this last is possible.

Claims

1. Process for purging a catalytic converter

(6) for treating the exhaust gases of an internal combustion engine (1), said pot (6) comprising means for adsorption of nitrogen oxides contained in these gases, according to which such a purge is triggered by an increase in the richness (R) of an air / fuel mixture supplying the engine, from a richness corresponding to a lean or stoichiometric mixture, characterized in that: a) one evaluates, using a table (Figure 2) depending on the temperature (T _cat ) and the filling rate (NS / NSC) of the pot (6) with nitrogen oxides, the efficiency (EFFNOX; EFFSTOCK) of the pot in terms of adsorption and reduction of nitrogen oxides, and b) a purge of the pot (6) is triggered when said efficiency (EFFNOX; EFFSTOCK) falls below a predetermined value (Eff_min), function of the temperature (T _ca t) of the pot (6).

2. Method according to claim 1, characterized in that said purging is stopped when the efficiency

(EFFNOX; EFFSTOCK) in lean mixture rises above a predetermined value (Eff_max) depending on the temperature (T _cat ) of the pot (6).

3. Method according to any one of claims 1 and 2, characterized in that the filling rate (NS / NSC) of the pot (6) is evaluated in nitrogen oxides, from an evaluation of the current quantity (NS) of nitrogen oxides stored during an operating phase of the engine in lean mixture, and the said rate (NS / NSC) is drawn from this evaluation and from a table (FIG. 3) giving the maximum quantity ( NSC) of nitrogen oxides adsorbable by the pot, depending on the temperature (T _cat ) of the pot and the current concentration of nitrogen oxides in the engine exhaust gas.

4. Method according to claim 3, characterized in that, for the evaluation of the current quantity (NS) of nitrogen oxides stored, the ratio (EFFDENOX) between the flow rate of oxides is taken into account nitrogen treated by the pot (6) and the flow of nitrogen oxides out of the engine, said ratio being taken from a table (Figure 6) depending on the temperature (T _cat ) of the pot (6).

5. Method according to claim 4, characterized in that, during a purge, continuously decreasing decreases of nitrogen oxides (DNSPURGE) destocked successively in the pot (6) and in that the purge is stopped when the efficiency (EFFNOX; EFFSTOCK) of the pot (6) rises above said threshold value (Eff_max) predetermined due to said successive destocking.

6. Method according to any one of claims 1 to 5, characterized in that said efficiency (EFFNOX) of the pot is measured by the ratio of the flow of nitrogen oxides at the outlet of an exhaust line (4χ , 4 ₂ ) of the engine (1) at the flow rate of said nitrogen oxides at the outlet of the engine (1).

7. Method according to any one of claims 4 and 5, characterized in that said efficiency is the storage efficiency (EFFSTOCK) such that: EFFSTOCK = (EFFNOX - EFFDENOX) 1 (1 - EFFDENOX).

8. Method according to any one of claims 6 and 7, characterized in that an instantaneous efficiency is evaluated.

9. Method according to any one of claims 6 and 7, characterized in that an average efficiency is evaluated.

10. Method according to any one of claims 1 to 9, characterized in that, during a tilting of the engine (1) of an operation in lean mixture to an operation at a richness greater than the stoichiometry, the richness is adjusted to a value (RICPURGE) corresponding to a purge command if the pot efficiency is then less than a threshold value (Eff_min) as a function of the temperature (T _cat ) of the pot (6).

11. Method according to any one of claims 1 to 10, characterized in that, during operation of the engine (1) at stoichiometry, the storage efficiency of the pot (6) is evaluated in the case of '' a possible operation to come in lean mixture and the richness of the mixture is adjusted to a value (RICPURGE) corresponding to a purge command if the pot efficiency (6) is then lower than a threshold value (Eff__min) function of the temperature (T _cat ) of the pot.

12. Method according to any one of claims 10 and 11, characterized in that the said purge richness (RICPURGE) is chosen from a table as a function of the pressure (P) at the manifold and of the engine speed (N) (1).

13. Method according to any one of the preceding claims, characterized in that, during a purging phase, the torque delivered by the motor (1) is maintained at a predetermined value.

14. Method according to claim 13, characterized in that, in order to maintain said torque, the increase in torque due to the rise in the richness of the mixture to the purge value (RICPURGE) is compensated for by an adapted reduction in ignition advance (Av).

15. Method according to claim 13, characterized in that, in order to maintain said torque, the increase in engine torque due to the rise in the richness of the mixture is compensated for at the purge value (RICPURGE), by an adapted reduction of the amount of air entering the engine (1).

16. Method according to claim 13, characterized in that, to maintain said torque, one compensates the increase in engine torque due to the increase in the richness of the mixture to the purge value (RICDNX), by a suitable reduction in the ignition advance (Av) and in the amount of air entering the engine .

17. Use of the method according to any one of claims 1 to 16, to control the purging of a catalytic converter (6) for treating the exhaust gases of an internal combustion engine (1) propelling a motor vehicle.