FR3067059A1 - PROCESS FOR OPTIMIZING DETERMINATION OF OXIDES OF NITROGEN FROM COMBUSTION GASES IN AN EXHAUST LINE - Google Patents

Abstract

The invention relates to a method of optimizing a nitrogen oxide pollution control of the combustion gases in an exhaust line (3), the depollution being carried out by at least two systems (1, 15; ) of selective catalytic reduction by injection by each system (1, 15; 2, 25) of an amount of reducing agent in the line (3) according to a quantity of agent instruction necessary to reduce the measured nitrogen oxides upstream of the system (1, 15; 2, 25), the injection being followed by a mixture of the amount of injected agent with the exhaust gas upstream of a catalyst (1, 2) of the system ( 1, 15; 2, 25). The two systems (1, 15; 2, 25) are successively arranged one after the other in the line (3) with, for each system (1, 15; 2, 25), an injection of a quantity of agent corresponding to the agent quantity setpoint decreased by a predetermined calibrated percentage before mixing.

Description

PROCESS FOR OPTIMIZING A NITROGEN OXIDE DEPOLLUTION OF GASES

OF COMBUSTION IN AN EXHAUST LINE [0001] The present invention relates to a method for optimizing a depollution of nitrogen oxides from combustion gases in an exhaust line of a heat engine, advantageously but not limited to a vehicle automotive, depollution of nitrogen oxides being carried out by two selective catalytic reduction systems. The present invention also relates to an exhaust line at the outlet of a heat engine of a motor vehicle for the implementation of such a method.

In the specific and non-limiting case of motor vehicles, compliance with current and future pollution control standards for motor vehicles with thermal engine implies elimination of the nitrogen oxides discharged from the engine in the exhaust gases by a line exhaust. Nitrogen oxides, hereinafter also referred to by their chemical formula NOx, are divided between nitrogen monoxide NO and nitrogen dioxide NO2 produced by combustion in the engine. The same will apply to ammonia, which can be designated by its chemical formula NH3.

This can be done by a selective catalytic reduction system. Such a Selective Catalytic Reduction system, known by the acronym RCS, is also known by the English acronym SCR. In what follows, reference will also be made to the complete denomination or to the abbreviation RCS to designate everything related to selective catalytic reduction.

Referring to Figure 1 which shows an exhaust line 3 equipped with an RCS system with representation of its injector 15 of reducing agent and its catalyst 1, an RCS system operates by injection into the line exhaust of a pollution control agent known as RCS reducer. This agent is advantageously but not limited to urea or a urea derivative, a precursor of ammonia which is used to reduce nitrogen oxides or NOx.

The reducing agent, illustrated in the form of black dots, is mixed in a mixing box not referenced in Figure 1 but located downstream of the injector 15 with the exhaust gases. The reduction of the nitrogen oxides, the nitrogen oxides being illustrated in the form of white dots, is carried out in the RCS catalyst 1 of the system downstream of the reducing agent injector 15 and the mixing box, l the reducing agent then being in the form of NH3. Downstream of the RCS 1 catalyst, there remain white dots representative of the nitrogen oxides which have decreased compared to those upstream. Ammonia may also remain, illustrated by black dots, obtained by transformation of the reducing agent, which is released into the environment. This means that the reduction of nitrogen oxides was not complete and that at the same time all the ammonia was not used during the reduction, due for example to an inhomogeneous mixing of NH3 in the exhaust gas.

At the outlet of the catalyst, NOx can be present in the gases because they are not treated with the reducing agent then in the form of NH3, this when the quantity of reducing agent injected is insufficient relative to the quantity of NOX to be reduced or NH3 may be present in the gases because it is not used for the reduction of NOx, this when the quantity of reducing agent injected is too large compared to the quantity of NOX to be reduced.

In Figure 1 it shows the remaining NOX, in the form of white dots as well as the remaining NH3, in the form of black dots, which is not the most frequent case, essentially corresponding to a poor mixture of the reducing agent with the exhaust gases. The most frequent cases are that there is either too much NOx at the outlet of the RCS 1 catalyst or either too much NH3, but not both, the "or" being the most frequently exclusive.

The quantity of reducing agent to be injected continuously is determined by an RCS control command from numerous elements including in particular the quantity of NOx entering the RCS system, the temperature of the exhaust gases and the flow rate of exhaust.

In the long term, estimating the mass of NH3 actually stored after transformation of the urea in the RCS catalyst is difficult and problematic. If the actual NH3 mass is less than the estimated mass, there is not enough injected reducing agent and there is a risk that the NOx will not be treated sufficiently. There will therefore be a surplus of NOx at the outlet of the RCS system catalyst.

If the actual NH3 mass is greater than the estimated mass, too much reducing agent has been injected and there is a risk of exceeding the maximum NH3 storage capacity of the catalyst of the RCS system. The NH3 in excess therefore not used for the reduction will be evacuated in the exhaust line downstream of the RCS system and released into the environment of the motor vehicle, if a catalyst for destruction of rejection or excess of NH3 is not provided at the outlet of the exhaust line.

This phenomenon is detrimental on two levels. At a first level, this reduces the autonomy of the vehicle as a reducing agent without gaining on the reduction efficiency of

NOx. At a second level, NH3 is released into the atmosphere, which either constitutes environmental pollution or either forces the exhaust line to be equipped with an ammonia release destruction catalyst, which increases the price and complexity of a motor vehicle exhaust line.

Depending on the precision of the NOx estimation or measurement upstream of the RCS injector and the precision of the reducing agent injector and the RCS catalyst taking into account the dispersions in operation, different situations are possible.

When the quantity of reducing agent precursor of NH3 actually injected is greater than the quantity of reducing agent sufficient to reduce the NOx actually present, the treatment efficiency is maximum but the excess NH3 is stored in the catalyst and participates in the NH3 mass drift in the RCS system.

When the quantity of reducing agent precursor of NH3 actually injected is less than the quantity of reducing agent sufficient to reduce the NOx actually present, the treatment efficiency is reduced and insufficient. If the missing NH3 is not compensated for by NH3 stored in the RCS catalyst, the NOx depollution is insufficient.

The optimal case is therefore the case where the amount of NH3 for the reduction of NOx actually present and the amount of NH3 released in the line after injection of an amount of reducing agent and, if necessary, the amount of NH3 present in the catalyst, correspond perfectly, which occurs little in practice. Different methods of taking these dispersions into account and of correcting them are possible, with more or less efficiency.

In addition, the non-control of the mass of NH3 available in the RCS system is problematic during temperature transients in the depollution line. When the vehicle accelerates, the line temperature rises sharply: the NH3 storage capacity of the RCS system then decreases very rapidly. If the quantity of NH3 stored exceeds this reduced storage capacity, the excess NH3 is desorbed and is released into the atmosphere.

In addition, in addition to the problems of precision of the NOx measured or estimated upstream of the RCS catalyst and of the reducing agent injector, other factors are likely to disturb the operation of the RCS system and limit its full effectiveness. .

The first factor is the sprinkling of the RCS catalyst with NOx and NH3, the NH3 having been mixed with the exhaust gases. However, this mixture is not perfect and inhomogeneous watering is observed in practice, which leads to preferentially sending NOx and NH3 to certain zones of the catalyst rather than others.

Frequently a large part of the NOx and NH3 is sent to the center of the RCS catalyst and the edges of the RCS catalyst are sparingly watered. This typically leads to treating the NOx well at the center of the RCS catalyst and to having reduced efficiency on the NOx passing around the periphery of the RCS catalyst.

The second factor is that most of the NOx sensors used measure NOx and NH3 without being able to distinguish the two: most NOx sensors have a dual sensitivity to NOX and NH3. Thus, in the event of excess NH3 unused, the NOx sensor downstream of the RCS system wrongly considers the surplus NH3 unused as NOx and considers that the NOx reduction efficiency is lower than in reality, so that the opposite happened: there was too much reducing agent injected giving a surplus of NH3, either stored in the RCS catalyst, or rejected in the exhaust line having not been used for NOx reduction.

EP-A-1 321 641 describes an exhaust gas emission purification system for a diesel engine comprising two diesel particle filters and two RCS catalysts. In addition to the first exhaust gas purification device, arranged in an engine exhaust outlet, comprising the first diesel diesel particulate filter with continuous regeneration and the first RCS catalyst, a derived exhaust outlet comprises a second exhaust device. exhaust gas purification comprising the second diesel particulate filter with continuous regeneration and the second RCS catalyst to constitute the exhaust gas purification system of the diesel engine.

In this document, the RCS catalysts are not aligned and the second catalyst is provided in a bypass. It follows that there is no correspondence between the downstream outlet of the first catalyst with the upstream inlet of the second catalyst and that the NH3 in surplus or the NOx remaining after the first RCS system are not necessarily treated in the second system. There is therefore no complementary cooperation between the two RCS systems to ensure effective depollution of NOx.

Therefore, the problem underlying the invention is, for a combustion gas exhaust line coming for example from a heat engine, this line comprising at least two RCS systems, to operate both systems so that the second system furthest downstream of the two systems can complete the reduction of NOx not carried out by the first system and use, if necessary, an excess of NH3 which has not been used for reduction of NOx in the catalyst

RCS of the first system, the cooperation of the two RCS systems leading to optimized depollution of NOx in the exhaust line.

To achieve this objective, there is provided according to the invention a method for optimizing a depollution of nitrogen oxides from the combustion gases in an exhaust line, the depollution of nitrogen oxides being carried out by at least two systems of selective catalytic reduction by injection by each system of a quantity of reducing agent in the line according to a set point of quantity of reducing agent necessary to reduce the nitrogen oxides estimated or measured upstream of the system, l injection being followed by a mixture of the quantity of reducing agent injected with the exhaust gases circulating in the exhaust line upstream of a reduction catalyst of the system, characterized in that the two systems are arranged successively one after the other in the line with, for each system, an injection of a quantity of reducing agent corresponding to the quantity reference of a reducer gent reduced by a predetermined percentage calibratable before mixing.

First of all, the present invention can be used for any type of heat engine, whether fixed or mobile, and not just for a motor vehicle engine, although this is a particularly well suited application.

What the present invention offers is not only the combination of two selective catalytic reduction systems called RCS systems. The method according to the present invention proposes to carry out an injection of reducing agent which is less than the setpoint for the quantity of reducing agent in order to have a deficit in NOx treatment rather than a surplus of NH3 which has not been used for the reduction.

Each of the catalysts of the first and second systems being in deficit of NH3 to treat all of the incoming NOx, the mass of NH3 stored in the catalysts is close to 0 permanently. Each NOx sensor therefore only measures NOx and cannot be disturbed by the presence of NH3 in the gases, which made it impossible, according to the state of the art, to differentiate a surplus of NH3 from an amount of untreated NOx and led to a runaway of the system in a bad direction of treatment after a false diagnosis, of the additional reducing agent being able to be injected wrongly into the system when one was already in surplus of NH3 not consumed for a catalytic reduction.

The present invention provides a method and an architecture for a pollution control line with two RCS systems making it possible to maximize the efficiency of treatment of NOx while limiting the risk of surplus NH3 not being used, this by placing itself in a position of NH3 deficiency for the treatment of NOx. As most of the NOx were treated by the first system, the deficit of NH3 for the second system is not too penalizing and few NOx are released into the line downstream of this second system because of this lack of NH3.

The sprinkling of a catalyst of an RCS system with a mixture of NOx and NH3 is not perfect: we observe inhomogeneous sprinkling which leads to preferentially send NOx and NH3 to certain areas of the catalyst rather than others. Redoing a mixture allows the NOx present in the exhaust gases to be properly reprocessed.

Advantageously, the setpoint for the quantity of reducing agent is reduced by a predetermined calibrated percentage of 10 to 20%. The calibration of the setpoint is used to keep possible dispersions in the measurements or estimates of NOx and possible dispersions of RCS systems.

Advantageously, the percentage can be calibrated, on the one hand, as a function of the dispersions of the measurements or estimates of nitrogen oxides upstream and downstream of the first system and, if necessary, upstream and downstream of the second system, and, on the other hand, as a function of the dispersions of the injectors and catalysts of the first and second systems, the dispersions of the measurements or estimates being evaluated without injection of reducing agent, the measurements or estimates having to be then equivalent upstream and downstream of the first system and, where appropriate, of the second system, and, after correction of the dispersions between upstream and downstream measurements or estimates, the dispersions of the injectors and of the catalysts being evaluated by injection of a substochiometric quantity of reducing agent for each system taken in isolation, this sub-stochiometric quantity having to correspond to a theoretical quantity of nitrogen oxides remaining downstream re spective of the first and second systems, a correction of the dispersions of the injectors and of the catalysts being carried out when a quantity of nitrogen oxides measured or estimated is not equal to the quantity of nitrogen oxides theoretical by updating the quantity of theoretical nitrogen oxides and of the setpoint for the quantity of reducing agent as a function of the quantity of theoretical nitrogen oxides.

It is difficult to determine between two upstream and downstream sensors which one or which are subject to negative or positive dispersion. The upstream and downstream sensors or NOX models should be recalibrated against each other. This is done without the injection of a reducing agent, it being understood that the catalysts do not contain a reserve of NH3: the upstream and downstream sensors must deliver the same values since there has been no reduction in NOx.

Similarly, to evaluate the dispersions of RCS systems, after recalibration of the sensors and models, a sub-stoichiometric injection of reducing agent will make it possible to compare the NOX actually remaining with the NOx which should theoretically remain and to recalibrate the setpoints d injection according to the dispersions of RCS systems.

The invention also relates to an assembly comprising a combustion gas exhaust line with at least first and second selective catalytic reduction systems and command and control units for each system, the line housing the catalysts therein. respective of the two systems and being crossed by respective first and second injectors of reducing agent upstream of the associated catalyst, the line or the control units incorporating means for determining a quantity of nitrogen upstream of the catalyst of each system, each control unit having means for determining a setpoint for the quantity of reducing agent to be injected into the line according to measurements or estimates for the means for determining a quantity of nitrogen, characterized in that the together implements such a process, the respective catalysts of the two systems being arranged in l one after the other, each control unit having means for memorizing a calibrated predetermined percentage of reduction, means for calculating the setpoint for the quantity of reducing agent with the subtracting of the predetermined percentage from the setpoint.

The principle underlying the present invention is to multiply the RCS catalysts and to inject on each of them less reducing agent transforming into NH3 than necessary to treat NOx. There can be more than two RCS systems online. The means for determining a quantity of nitrogen can be sensors integrated in the exhaust line or models of nitrogen oxides integrated in the control and command units of RCS systems.

Advantageously, the assembly comprises from upstream to downstream a first nitrogen oxide sensor, a first injector, a first mixer and a first catalyst of the first system, a second nitrogen oxide sensor, a second injector, a second mixer and a second catalyst of the second system. The advantage of this solution is its reduced price but this solution does not allow the measurement of nitrogen oxides at the outlet of the exhaust line downstream of the second RCS system.

A reduction in the surplus of unused NH3 is obtained, which leads to a risk of environmental pollution and a reduction in the autonomy of reducing agent while ensuring optimal and robust treatment of the dispersions of the NOx sensors and reducing agent injectors.

Advantageously, a third nitrogen oxide sensor is positioned downstream of the second catalyst of the second system. This solution is very effective but costs more than the previous one by the presence of a third nitrogen oxide sensor at the outlet of the exhaust line.

Advantageously, the assembly comprises from upstream to downstream a model of nitrogen oxides for an estimation of the quantity of nitrogen oxides upstream of the first system comprising a first injector, a first mixer and a first catalyst, a first nitrogen oxide sensor interposed between first and second systems, the second system comprising a second injector, a second mixer and a second catalyst.

This solution corresponds to the previous one, however with the replacement of the first nitrogen oxide sensor with a nitrogen oxide model and the removal of the third nitrogen oxide sensor downstream from the second RCS system. The presence of a nitrogen oxide model in place of a nitrogen oxide sensor lowers the price of the exhaust line.

Advantageously, a second nitrogen oxide sensor is positioned downstream of the second catalyst. This solution is very effective but costs more than the previous one by the presence of a nitrogen oxide sensor at the outlet of the exhaust line.

Advantageously, the assembly comprises from upstream to downstream a first nitrogen oxide sensor, a first injector, a first mixer and a first catalyst of the first system, a model of nitrogen oxides for an estimate. the amount of nitrogen oxides downstream of the first system, a second injector, a second mixer and a second catalyst of the second system, a second nitrogen oxide sensor being positioned downstream of the second catalyst.

In this embodiment, the nitrogen oxide model replaces the second nitrogen oxide sensor. This embodiment makes it possible to validate the system during a diagnosis by an on-board unit.

Advantageously, the assembly comprises at least one of the following elements: an oxidation catalyst, at least one passive or active nitrogen oxide sensor disposed upstream of the first selective catalytic reduction system, a particle filter and an ammonia rejection catalyst disposed downstream of the second selective catalytic reduction system.

Other characteristics, aims and advantages of the present invention will appear on reading the detailed description which follows and with regard to the appended drawings given by way of nonlimiting examples and in which:

FIG. 1 is a schematic representation of an exhaust line comprising an RCS system according to the state of the art,

FIG. 2 shows two efficiency curves of two successive RCS systems online as a function of the cumulative population, this FIG. 2 showing the efficiency of a line equipped with two RCS systems operating in accordance with a method according to the present invention,

- Figures 3 to 7 illustrate respective embodiments of an exhaust system with two RCS systems according to the present invention.

It should be borne in mind that the figures are given by way of examples and are not limitative of the invention. They constitute schematic representations of principle intended to facilitate the understanding of the invention and are not necessarily at the scale of practical applications. In particular, the dimensions of the various elements illustrated are not representative of reality.

In what follows, reference is made to all the figures taken in combination. When reference is made to one or more specific figures, these figures are to be taken in combination with the other figures for the recognition of the designated numerical references. Upstream and downstream are to be taken by considering the path of the exhaust gases in the exhaust line.

Referring to Figures 2 to 7, the present invention relates to a method for optimizing a depollution of nitrogen oxides from the gases in a heat engine exhaust line 3, advantageously but not exclusively for a vehicle automobile. In this process, the depollution of nitrogen oxides is carried out by at least two systems 1, 15; 2, 25 of selective catalytic reduction.

Conventionally, for each RCS system 1, 15; 2, 25 an injection is carried out by an injector 15, 25 of a quantity of reducing agent in line 3 according to a setpoint of quantity of reducing agent necessary to reduce the nitrogen oxides estimated or measured upstream of the system . This injection is followed by a mixture of the quantity of reducing agent injected with the exhaust gases circulating in the exhaust line 3 upstream of a catalyst 1,2 for reducing the system 1, 15; 2, 25.

According to the invention, the two RCS systems 1, 15; 2, 25 are arranged successively one after the other in line 3 and interact together with, for each system 1, 15; 2, 25, an injection of a quantity of reducing agent corresponding to the setpoint of quantity of reducing agent reduced by a predetermined percentage which can be calibrated before mixing.

After the first RCS 1.15 system, as the quantity injected was less than the setpoint for the quantity of reducing agent, all the NOx present in the gases were not reduced. In addition, as previously indicated, the first catalyst 1 treats the mixture better at its center than at its periphery.

Thanks to the second RCS 2, 25 system, there is a second injection of reducing agent and a new mixture of reducing agent injected with the exhaust gases. This will allow a new distribution of the reducing agent in the exhaust gases and a better elimination of NOx. The quantity of reducing agent injected for the second SCR 2, 25 system is also reduced compared to the setpoint, but the second SCR 2, 25 system can benefit from the NH3 stored in the first catalyst 1 and from a new mixture.

The estimation of the NO2 / NOx ratio upstream of the RCS is necessary to estimate the quantity of reducing agent to be injected since the different NOx reduction reactions have specific stoichiometries: for some it is necessary to inject as much NH3 as NOx, others require more NH3 than NOx.

As previously indicated, the urea-reducing reducing agent for ammonia, based on urea, the most widely used of which is known by the name AdBlue®, reacts at a high temperature to become ammonia or NH3 , NH3 reacting with nitrogen oxides or NOx, mainly in the form of a mixture of nitrogen monoxide or NO and nitrogen dioxide or NO2 with a ratio varying in particular according to the operating conditions of the engine and the temperature in exhaust line 3.

The depollution treatment of another pollutant or the maintenance of another depollution element, for example a regeneration of a particulate filter, can also interact on the NO / NO2 ratio.

The decomposition of urea into NH3 follows the following equation:

CO NH _{2 2} + H ₂ O 2 NH ₃ + CO ₂

This applies to mixtures based on urea as a reducing agent which is a precursor to ammonia.

For the reduction of NOx, NH3 in turn reacts with nitrogen oxides to form, by a reduction reaction, dinitrogen and water. For example, with nitrogen monoxide, the reaction is written:

NO + 4 NH ₃ + O ₂ 4 N ₂ + 6 H ₂ O

Another reaction with nitrogen monoxide and nitrogen dioxide is written:

NO + 4 NH ₃ + 2NO ₂ 4 N ₂ + 6 H ₂ O

Other chemical reactions between NOx and NH3 are also possible.

For an exhaust line 3 comprising first and second systems 1, 15; 2, 25 of selective catalytic reduction arranged one after the other, the ammonia not used by the first RCS system 1, 15 for the reduction of NOx can reach the second RCS system 2, 25. In the nonlimiting case of 'A motor vehicle, this can in particular occur during an increase in temperature in the exhaust line 3, for which ammonia is desorbed. Such an increase in temperature in the exhaust line 3 can occur during pronounced acceleration of the vehicle or when stabilized at high engine speed.

It is possible that the release of ammonia is greater than necessary to allow the reduction of NOx by the second RCS system 2, 25. In this case, there remains a surplus known as NH3 leakage which is evacuated in the environment by leaving the exhaust line 3, when the increase in temperature affects the second system 2, 25. Since an NH3 emission is a toxic emission, it is advisable to neutralize or prevent the creation of such an NH3 leak .

To do this, the solution of the present invention provides for injecting a quantity of reducing agent in the exhaust line 3 which is less than the setpoint for the quantity of reducing agent to be injected, this for the two RCS systems. 1.15; 2, 25.

Each of the catalysts 1, 2 being in deficit of NH3 to treat all of the incoming NOx, the mass of NH3 stored is close to 0 permanently. Each NOx sensor therefore only measures NOx and cannot be disturbed by the presence of NH3 in the gases.

FIG. 2, while referring to FIGS. 2 to 7 for the missing references in FIG. 2, shows the advantages provided by the present invention by showing on the ordinate an efficiency of treatment of the nitrogen oxides Eff DENOX and in abscissa a cumulative Pop cum population, a percentage of the number of escape routes. The curve in solid line is the curve leaving the first system 1.15 RCS while the curve with a line carrying circles is the curve leaving the second system 2, 25.

The lower NOx treatment efficiency at the periphery of the catalysts 1, 2 is partially reduced: the untreated NOx at the periphery of the catalyst 1 are remixed with reducing agent between the two catalysts 1,2 RCS and are distributed over the entire inlet surface of the catalyst 2. The efficiency will remain lower on the few NOx remaining at the periphery of the second catalyst 2 of the second system 2, 25, but this amount is marginal.

For example, it is assumed that for each measurement of NOx and for each injector 15, 25 of reducing agent, the precision follows a normal law at +/- 10%, that is to say that 100 NOx measured correspond to reality between 90 and 110 of NOx actually present and that 100 NH3 requested correspond between 90 and 110 NH3 actually injected in reality. Indeed, it is not apparent in which direction the dispersions concerning the NOx measurements or estimates go and the dispersions concerning the first and second injectors 15, 25 and the first and second 2 SCR catalysts, these dispersions being able to be negative as well than positive.

Still referring to FIG. 2 and to one of FIGS. 3 to 7, in order to achieve maximum efficiency while remaining NH3 deficient on the first catalyst 1 of the first RCS system 1, 15, it is appropriate to aim for efficiency around 80% nominal. Given the dispersions of the first sensor 14 and the first injector 15, the efficiency of the first RCS system 1, 15 is shown in the lower curve in solid lines of the two curves illustrated in FIG. 2.

NOx pollution control in the exhaust line 3 with a first 14 NOx sensor upstream overestimating the real NOx, for example in the extreme 90 real NOx per 100 measured, and a first injector 15 over-discharging, by example in the extreme 90 NH3 injected for 80 requested, with a target efficiency of 80% on 100 NOx measured, reaches 100% of real efficiency.

NOx depollution in the exhaust line 3 with a first 14 NOx sensor upstream underestimating the real NOx, for example to the extreme 110 real NOx per 100 measured and a first sub-flow injector 15, for example in the extreme 70 NH3 injected for 80 requested, with a target efficiency of 80% on 100 NOx measured, reaches 64% of real efficiency by being the ratio 70/110.

The rest of the population is between these two extremes: due to the dispersions, 59% of the population exceeds 80% of efficiency on the first catalyst 1 of the first RCS system 1, 15 and 8% exceeds 90% of efficiency on the first catalyst 1 of the first RCS system 1.15.

NOx remaining after the first catalyst 1 RCS of the first system 1.15. According to the invention, there are only NOx and no NH3, since the quantity of reducing agent injected is substoichiometric by being below the setpoint. The quantity of NOx remaining after the first RCS system 1, 15 is again measured via a second sensor 24 downstream of the first RCS system 1, 15 and upstream of the second RCS system 2, 25. The quantity to be measured being smaller, the impact of dispersions is reduced. For example, by measuring 10 NOx downstream of the first RCS system 1.15, the actual NOx is between 9 and 11. For a request to inject 10 NH3, it is injected between 9 and 11 NH3.

It is then established the same dispersion study as previously for the second RCS system 2, 25. As shown in the top curve of Figure 2 having circles, the impact of dispersions is significantly reduced: 99% of exhaust lines 3 exceed 90% of actual efficiency and 84% of exhaust lines 3 exceed 95% efficiency.

The solution of the present invention makes it possible to reduce the impact of the estimation of the NO2 / NOx ratio upstream of the first and second RCS systems 1, 15; 2, 25. According to this value, the injection must be modulated to take into account the stoichiometry of the various NOx treatment reactions. However, this estimate can be a source of imprecision on the injection to be requested, and ultimately on the efficiency of the system.

For example, the NO2 / NOx ratio upstream of the first estimated RCS system 1.15 is 0.5, therefore the stoichiometry of the NOx reduction reactions is 1 NOx / 1 NH3. If you have to treat 100 NOx, i.e. 50 NO2 and 50 NO, taking into account the stoichiometry, inject 100 NH3 and expect to treat everything.

The NO2 / NOx ratio upstream of the first real RCS system 1.15 is 0.6, therefore the stoichiometry of the NOx reduction reactions is 1 NOx for 1.1 NH3. As it is necessary to treat 100 NOx, that is to say 60 NO2 and 40 NO, taking into account the stoichiometry one should have asked for an injection of 110 NH3 instead of 100 NH3 by estimating the ratio. We will therefore process 10 NOx less than expected. These untreated NOx will be measured by the second sensor 24 downstream of the first RCS system 1, 15 and upstream of the second RCS system 2, 25 and more injection into the second RCS system 2, 25 will be requested on the second catalyst 2 to treat them.

The quantity of reducing agent can be reduced by a predetermined calibrated percentage of 10 to 20%. This percentage can be calibrated to take account of the various dispersions, on the one hand, the dispersions on the measurement or the estimation of NOx and, on the other hand, the dispersions in each RCS system 1, 15; 2, 25, for example the dispersions of the first and second injectors 15, 25 and of the first and second catalysts 1,2.

The percentage can therefore be calibrated, on the one hand, as a function of the dispersions of the measurements or estimates of nitrogen oxides upstream and downstream of the first system 1, 15 and, if necessary, upstream and downstream of the second system 2, 25, and, on the other hand, as a function of the dispersions of the injectors 15, 25 and of the catalysts 1, 2 of the first and second RCS systems 1, 15; 2, 25. Examples of correcting dispersions of measurements or estimates, on the one hand, and dispersions of RCS systems 1, 15; 2, 25, on the other hand are going to be given. These examples are not limiting but purely illustrative.

If it is difficult to determine which of two NOx sensors give erroneous values, it is possible to calibrate two NOx sensors respectively upstream and downstream of an RCS system with respect to each other. The dispersions of the measurements or estimates can be evaluated without injecting a reducing agent. In this case no NOx reduction takes place and the measurements or estimates must then be equivalent upstream and downstream of the first system 1, 15 and, if necessary, of the second system 2, 25. It may be possible to calibrate the upstream and downstream sensors on the sensor giving the greatest or the lowest NOx value, advantageously the greatest value. This is done with catalysts 1, 2 which do not contain a reserve of NH3.

After this correction of the dispersions between upstream and downstream measurements or estimates, the dispersions of the injectors 15, 25 and of the catalysts 1, 2 can be evaluated by injection of a sub-stochiometric quantity of reducing agent for each system 1, 15; 2, 25 taken in isolation. This substoichiometric quantity having to correspond to a theoretical quantity of nitrogen oxides remaining downstream of the respective first and second systems 1, 15; 2, 25.

The actual amount of nitrogen oxides remaining is measured or estimated with corrected measurement or estimation means and a correction of the dispersions of the injectors 15, 25 and of the catalysts 1, 2 can be carried out when a quantity d The measured or estimated nitrogen oxides is not equal to the theoretical quantity of nitrogen oxides. This can be done by updating the quantity of theoretical nitrogen oxides and the instruction for the quantity of reducing agent as a function of the quantity of theoretical nitrogen oxides.

The invention also relates to a line 3 for escaping gases from combustion in a heat engine, advantageously of a motor vehicle, with at least first and second systems 1, 15; 2, 25 of selective catalytic reduction and control units of each of the first and second systems 1, 15; 2, 25, the line and the control units forming a unit. Line 3 of exhaust houses inside its first and second catalysts 1, 2 respectively of the two systems 1, 15; 2, 25 and is crossed by the first and second injectors 15, 25 respectively of reducing agent upstream of the associated catalyst 1,2.

Line 3 incorporates a nitrogen oxide sensor 14 or a model 16 of nitrogen oxides estimating a quantity of nitrogen oxides upstream of the catalyst 1, 2 of each system 1, 15; 2, 25, the sensor 14 and the model 16 forming means for determining the quantity of nitrogen previously mentioned.

Each system 1, 15; 2, 25 comprises a control unit presenting means for determining a setpoint for the quantity of reducing agent to be injected into line 3 according to the measurements of the sensor (s) 14 or the estimates of the model (s) of oxides d 'nitrogen. The two command and control units can be combined into a single command and control unit common to the two RCS systems 1, 15; 2, 25.

According to the invention, the assembly implements such a process with the respective catalysts 1, 2 of the two systems 1, 15; 2, 25 arranged in line one after the other, all the flow of exhaust gas leaving the first RCS system 1.15 passing through the second RCS system 2, 25. The control unit has means of memorization of a calibrated predetermined percentage of reduction, means for calculating the setpoint for the quantity of reducing agent with deduction of the predetermined percentage from the setpoint.

One can choose different RCS technologies for each catalyst. For example, we can only use technologies based on iron, copper, or a mixture of the two. In particular, an iron catalyst may be used for the first catalyst 1 which operates at a lower temperature and a copper catalyst for the second because such a catalyst is less sensitive to the NO2 / NOx ratio, an unknown value downstream of the first catalyst. 1 of the first RCS system 1.15.

Different types of substrates for catalysts 1, 2 may be possible: either a conventional cordierite catalyst substrate already used for many 1,2 catalysts such as a diesel oxidation catalyst or a passive NOx trap, or a silicon carbide filter substrate as for a particle or cordierite filter, as for a particle filter suitable for petrol or aluminum titanate engines, on which an RCS impregnation is placed.

In the context of the invention, there may be more than two RCS systems 1, 15; 2.25 arranged one after the other. A NOx 14 upstream sensor or a NOx 16 model measures or estimates the NOx to be treated according to the operating conditions, essentially a motor point, an ambient temperature, a hygrometry, etc.

The first and second injectors 15, 25 of reducing agent respectively upstream of the first and second catalysts 1,2 provide the reducing agent to give the necessary NH3 according to the amount of NOx to be treated and an evaluated NO2 / NOx ratio reflecting the proportion of NO2 among the NOx.

The first and second mixers or mixers make it possible to homogenize the mixture of reducing agent and exhaust gas between, on the one hand, the first and second injectors 15, 25 and, on the other hand, respectively the first and second catalysts 1, 2 RCS.

A downstream NOx sensor 24 is added for the first system 1, 15 which is the upstream NOX sensor for the second system 2, 25 to evaluate the efficiency of NOx treatment and to verify that each of the first and second RCS systems 1.15; 2, 25 is working properly. A NOx sensor downstream 34 of the second RCS system 2, 25 can also be provided.

In general, it can be selected between a NOx sensor 14 and a NOx model 16 for measuring or estimating NOx upstream of one of the first and second RCS systems 1, 15; 2, 25. Non-limiting examples will now be given with reference to FIGS. 3 to 7. These examples are not limiting and other examples can be used without departing from the scope of the present invention. For all these embodiments, there remain quantities of NOx downstream of the first RCS system 1.15 reduced compared to the quantity of NOx upstream of the first RCS system 1,

15.

A quantity of reducing agent is also injected upstream of the second system

SCR 2, 25. These quantities of NOx and NH3 are symbolized respectively by white dots and black dots. Downstream of the second RCS 2 system, 25 there remains no NH3 and very little NOx, which is an indication of an optimal depollution of NOx without rejection of an excess of NH3 not used during the catalytic reduction.

Referring to Figure 3, the exhaust line 3 may comprise from upstream to downstream a first sensor 14 of nitrogen oxides, a first injector 15, a first mixer and a first catalyst 1 of the first system 1.15. Can then follow a second sensor 24 of nitrogen oxides, a second injector 25, a second mixer and a second catalyst 2 of the second system 2, 25.

Referring to Figure 4, the arrangement of the exhaust line 3 of Figure 3 can be repeated with a third sensor 34 of nitrogen oxides which is positioned downstream of the second catalyst 2 of the second system 2 , 25.

With reference to FIG. 5, the exhaust line 3 may comprise from upstream to downstream a model 16 of nitrogen oxides for an estimation of the quantity of nitrogen oxides upstream of the first system 1 , 15 comprising a first injector 15, a first mixer and a first catalyst 1. A first nitrogen oxide sensor 14 is interposed between the first and second systems 1, 15; 2, 25, the second system 2, 25 comprising a second injector 25, a second mixer and a second catalyst

2.

Referring to Figure 6, the arrangement of the exhaust line 3 of Figure 5 can be resumed and a second sensor 34 of nitrogen oxides can be positioned downstream of the second catalyst 2.

Referring to Figure 7, the exhaust line 3 may comprise from upstream to downstream a first sensor 14 of nitrogen oxides, a first injector 15, a first mixer and a first catalyst 1 of the first system 1.15. Then there is provided a model 26 of nitrogen oxides for an estimation of the quantity of nitrogen oxides downstream of the first system 1, 15, a second injector 25, a second mixer and a second catalyst 2 of the second system 2, 25, a second nitrogen oxide sensor 34 being positioned downstream of the second catalyst 2.

In addition to the above elements, the exhaust line 3 may include at least one of the following pollution control elements: an oxidation catalyst, at least one passive or active nitrogen oxide sensor disposed upstream of the first selective catalytic reduction system 1, 15, a particulate filter and an ammonia rejection catalyst disposed downstream of the second selective catalytic reduction system 2, 25.

Some of these elements can be integrated between the first and second catalysts 1, 2. An ammonia rejection catalyst, also called Clean Up Catalyst or Ammonia Slip Catalyst in the English language, eliminates the surplus of unused NH3 for the selective catalytic reduction of the two RCS systems 1, 15; 2, 25 consecutive in line 3 of exhaust. The ammonia release catalyst is located downstream of the two RCS systems 1, 15; 2, 25 in the exhaust line 3, advantageously in the downstream end portion of the exhaust line 3.

The exhaust line 3 may also include pressure sensors, in particular at the terminals of a particle filter, at least one oxygen sensor and a soot sensor. This is not limiting.

During cold starts, the first catalyst 1 will heat up and therefore activate more quickly than the second 2, an injection of liquid reducing agent being effective only from a certain temperature of the gases d 'exhaust. In this case, we can choose to work only with the first catalyst 1 and do all of the NOx reduction on it, as long as the second RCS system 2, 25 is not activated and qu 'the second injector 25 cannot be used.

When the entire exhaust line 3 is hot enough, one can start working with the second catalyst 2 of the second RCS 2 system 25, in accordance with the method described above.

The situation is identical during the phases of transition from slow travel to rapid travel. Conversely, during the transition phases from fast running to slow running, the temperature at the engine outlet will successively decrease and cool the various catalysts 1, 2. Thus, while the first catalyst 1 RCS will cool and deactivate, the second catalyst 2 will stay hot longer and can temporarily ensure the total reduction of NOx before deactivating in turn if slow driving is prolonged.

The possible architectures of an exhaust line 3 in accordance with the present invention are more resistant to thermal sweeping in the event of strong accelerations. As the mass of NH3 stored in catalysts 1, 2 is low, or even zero, there is nothing to destock in the event of strong acceleration and drop in storage capacity.

The invention is in no way limited to the embodiments described and illustrated which have been given only by way of examples.

Claims (10)

1. Method for optimizing depollution of nitrogen oxides from combustion gases in an exhaust line (3), depollution of nitrogen oxides being carried out by at least two systems (1, 15; 2, 25) of selective catalytic reduction by injection by each system (1, 15; 2, 25) of a quantity of reducing agent in line (3) according to a setpoint of quantity of reducing agent necessary to reduce the oxides of nitrogen estimated or measured upstream of the system (1, 15; 2, 25), the injection being followed by a mixture of the quantity of reducing agent injected with the exhaust gases circulating in the line (3) of exhaust upstream of a catalyst (1, 2) reduction of the system (1, 15; 2, 25), characterized in that the two systems (1, 15; 2, 25) are arranged successively one after the other in the line (3) with, for each system (1, 15; 2, 25), an injection of a quantity of reducing agent corresponding to the setpoint of quantity of reducing agent reduced by a predetermined percentage which can be calibrated before mixing. 2. The method of claim 1, wherein the quantity of reducing agent setpoint is reduced by a predetermined calibrated percentage of 10 to 20%. 3. Method according to claim 2, in which the percentage can be calibrated, on the one hand, as a function of the dispersions of the measurements or estimates of nitrogen oxides upstream and downstream of the first system (1, 15) and, the if necessary, upstream and downstream of the second system (2, 25), and, on the other hand, as a function of the dispersions of the injectors (15, 25) and of the catalysts (1, 2) of the first and second systems (1 , 15; 2, 25), the dispersions of the measurements or estimates being evaluated without injecting a reducing agent, the measurements or estimates then having to be equivalent upstream and downstream of the first system (1, 15) and, if necessary, of the second system (2, 25), and, after correction of the dispersions between upstream and downstream measurements or estimates, the dispersions of the injectors (15, 25) and of the catalysts (1, 2) being evaluated by injection of an under- stochiometric of reducing agent for each system (1, 15; 2, 25) taken in isolation, this e sub-stochiometric quantity which must correspond to a theoretical quantity of nitrogen oxides remaining downstream of the first and second systems respectively (1, 15; 2, 25), a correction of the dispersions of the injectors (15, 25) and of the catalysts (1, 2) being carried out when a quantity of nitrogen oxides measured or estimated is not equal to the quantity of oxides d theoretical nitrogen by updating the quantity of theoretical nitrogen oxides and the setpoint of quantity of reducing agent as a function of the quantity of theoretical nitrogen oxides. 4. Assembly comprising a line (3) of combustion gas exhaust with at least first and second systems (1, 15; 2, 25) of selective catalytic reduction and control and command units of each system, the line ( 3) housing in its interior the respective catalysts (1,2) of the two systems (1, 15; 2, 25) and being crossed by respective first and second injectors (15, 25) of reducing agent upstream of the associated catalyst (1, 2), the control units incorporating means for determining an amount of nitrogen upstream of the catalyst (1, 2) of each system (1, 15; 2, 25), each control unit having means for determining a setpoint for the quantity of reducing agent to be injected into the line (3 ) according to the measurements of the sensor (s) (14) or the estimates of the model (s) (16) of nitrogen oxides, characterized in that the assembly implements a method according to any one of the preceding claims, respective catalysts (1, 2) of the two systems (1, 15; 2, 25) being arranged in line one after the other, each control unit having means for memorizing a calibrated predetermined percentage of decrease, means for calculating the setpoint of quantity of reducing agent, subtracting from the predetermined percentage of the set point. 5. The assembly of claim 4, wherein the line (3) comprises from upstream to downstream a first sensor (14) of nitrogen oxides, a first injector (15), a first mixer and a first catalyst (1 ) of the first system (1, 15), a second sensor (24) of nitrogen oxides, a second injector (25), a second mixer and a second catalyst (2) of the second system (2, 25). 6. The assembly of claim 5, wherein a third sensor (34) of nitrogen oxides is positioned downstream of the second catalyst (2) of the second system (2, 25). 7. The assembly of claim 4, wherein the line (3) comprises from upstream to downstream a model (16) of nitrogen oxides for an estimation of the amount of nitrogen oxides upstream of the first system ( 1, 15) comprising a first injector (15), a first mixer and a first catalyst (1), a first sensor (24) of nitrogen oxides interposed between first and second systems (1, 15; 2, 25 ), the second system (2, 25) comprising a second injector (25), a second mixer and a second catalyst (2). 8. Assembly according to the preceding claim, in which a second sensor (34) of nitrogen oxides is positioned downstream of the second catalyst (2). 9. The assembly of claim 4, wherein the line (3) comprises from upstream to downstream a first sensor (14) of nitrogen oxides, a first injector (15), a first mixer and a first catalyst (1 ) of the first system (1, 15), a model (26) of nitrogen oxides for an estimation of the quantity of nitrogen oxides in 5 downstream of the first system (1, 15), a second injector (25), a second mixer and a second catalyst (2) of the second system (2, 25), a second sensor (34) of nitrogen oxides being positioned downstream of the second catalyst (2). 10. Assembly according to any one of claims 4 to 9, in which the line (3) comprises at least one of the following elements: an oxidation catalyst, at 10 minus a passive or active nitrogen oxide sensor disposed upstream of the first selective catalytic reduction system (1, 15), a particle filter and an ammonia rejection catalyst disposed downstream of the second system (2, 25) selective catalytic reduction.