FR2927947A1 - Sulfur oxide purging control method for e.g. diesel engine of motor vehicle, involves determining rich time/weak time such that rich time is higher than duration necessary for consuming surface oxygen present in converter - Google Patents

Abstract

The method involves evaluating a variation of mass of purged sulfur oxide in a catalytic converter (4), by a unit of time based on temperature of the converter, richness of air/fuel mixture and stored oxygen mass capacity that considers quantity or presence of surface oxygen. Determination is made that desorption of sulfur is carried out only when the richness is higher than 1 or in the absence of the surface oxygen. A rich time/weak time is determined such that the rich time is higher than a duration necessary for consuming the surface oxygen present in the converter.

Description

The present invention relates to a method for purging sulfur oxides of a catalytic converter for treating the exhaust gases of an internal combustion engine. In order to meet the lower thresholds for emissions of gaseous pollutants from motor vehicles, catalytic systems, of increasingly complex gas after-treatment, are arranged in the exhaust line of lean-burn engines. These reduce particulate and nitrogen oxide emissions, in addition to carbon monoxide and unburned hydrocarbons. Unlike conventional oxidation catalysts, these systems operate discontinuously or alternatively, that is, in normal operation, they trap pollutants, but only treat them during regeneration phases.

Thus, to be regenerated, these traps require specific modes of combustion to ensure the necessary thermal and / or wealth levels. The storage capacity of a catalyst comprising a nitrogen oxide trap (or NOx) is altered over time by the presence of sulfur in the exhaust gas. As it is indeed present in the fuel and the oil, it is found in the exhaust. It is trapped on the reaction sites of the trap, thus decreasing the number of reactive sites for NOx. In other words, the efficiency of a NOx trap decreases with the accumulation of sulfur in it.

As for NOx, it is necessary after having crossed a certain loading threshold, to eliminate (or desorb sulfur) to restore the trap of the efficiency in terms of NOx absorption. This process, called desulfation or deSOx, requires a reducing medium (rich mixture) like the NOx purges, but also a high thermal level in the trap (NOx-Trap) (of the order of 700 ° C). This operation has impacts in terms of mechanical strength of the engine (extreme thermal of the various organs), fuel dilution in oil and consumption. In addition, the operating conditions are difficult to obtain on the vehicle. The main difficulty of desulfation is to keep the temperature of the NOxtrap below a maximum admissible temperature to avoid its destruction, while maintaining a sufficient thermal level to make possible the desorption of SOx. To do this, the control of desulphatation is carried out by maintaining the temperature with a flow of rich gas continuously or by an alternation of richness of the mixture at the entrance of the NOxtrap around the wealth 1. The poor phase (lower wealth to 1) allows then to reduce the temperature of the NOxtrap (by temporary stop of the desulfation) and to reload the oxygen of surface (abbreviated OS, ie oxygen stored on the trap) of the NOxtrap by the oxygen contents in the exhaust gases (the emission downstream of the S02 sulfur dioxide trap or the H2S hydrogen sulphide trap depends on the amount of OS). The rich phase makes it possible to achieve the effective deSOx under certain conditions. It has already been proposed in EP-A-0 962 639 to estimate the sulfur poisoning of a catalyst.

This sulfur oxide (SOx) purge control method of a catalytic converter for treating the exhaust gases of an internal combustion engine, said pot comprising means for adsorbing sulfur oxides (SOx) contained therein in these gases, comprises the following steps: a) the current mass (SSi) of sulfur oxides adsorbed in the pot at a sampling instant i is evaluated, using the model: SSi + 1 = SSi + Qsox. EFFSOX where Qsox is the flow rate of sulfur oxides in the exhaust 30 of the engine and EFFSOX is the instantaneous storage efficiency of the sulfur oxides in the pot, b) a purge of the pot is triggered when said quantity crosses a threshold ( Si) predetermined, a function of the temperature (Tcat) of the pot, by an increase in the richness (Ri) of an air / fuel mixture 35 of supply of the engine, from a richness corresponding to a lean or stoichiometric mixture , c) alternating injections of a rich mixture (Ri> 1) and poor (Ri <_1), for respective durations called "rich time" and "poor time". Richness fractionation (ie the alternation of periods during which the mixture is rich then poor) is a strategy particularly used in the control of the desulfation of a NOx Trap, to control the temperature and the nature. sulfur emissions downstream of NOxTRAP. It is a strategy that induces many constraints in its management by the difficulty to estimate the moment of the best passage from a rich mixture to a poor mixture. Thus, all that makes it possible to better apprehend the conditions of changeover from one to the other brings an advantage in the control of the strategy. The duration of the rich and poor alternating phases during desulfation is currently predetermined by calibration or is the consequence of a temperature control strategy. Calibration of "rich" and "poor" times (or guidelines of the temperature control strategy) is defined to optimize desulfation rate as well as SO2 emission, rather than H2S. Since these properties are intrinsically linked to the capacity of the oxygen storage trap (OS), and the latter evolving with the thermal aging and the sulfur poisoning of the trap, the current strategy does not know how to adapt with the evolution of NOxTRAP. It is proposed by the present invention to optimize the speed and the desulfation costs (dilution, overconsumption, etc.), as well as the selectivity of downstream sulfur species (SO2 rather than H2S), by modifying the management of the splitting of wealth. This invention is based on the assumption that sulfur desorption takes place in SO2, H2S and COS form in rich medium, once the surface oxygen (OS) of the trap is fully consumed by the excess reductants. It is therefore a process as described above, which is essentially remarkable in that: d) in step b), the variation of the oxide mass is evaluated for said pot; of sulfur (SOx) purged per unit of time dmSOx / dt, in particular as a function of the temperature of the pot (Tcat), the richness (Ri) of the air / fuel mixture, as well as a quantity "mO2 stored "which takes into account the quantity or the presence of surface oxygen (OS), this" surface oxygen "(OS) being the oxygen impregnated on the surface of the adsorption means of said pot, e) it is considered that the desorption of sulfur takes place only when said richness (Ri) is greater than 1 and in the absence of surface oxygen (OS); f) the rich time (tr) / lean time (tp) pair is determined in such a way that the rich time (tr) is greater than the time required to consume the surface oxygen (OS) present in the pot. Taking into account the "surface oxygen" parameter makes it possible to optimize fractionation (triggering and duration of rich and poor niches) during desulfation. It is thus possible to reduce the production of malodorous species (ie to favor the formation of SO2 to the formation of H2S), to reduce the time of desulfation (by increasing the efficiency of desulfation) and to reduce the overconsumption and the fuel dilution induced by the strategy. According to advantageous and nonlimiting features of this method: the determination of said rich time (tr) is carried out using means for detecting the presence or absence of surface oxygen (OS); said detection means consist of at least one oxygen probe; said detection means consist of a pattern of consumption of surface oxygen (OS); the determination of said rich time (tr) is greater than the consumption time of the surface oxygen (OS) and corresponds to the time for which a predetermined threshold of desorbed sulfur is reached; the said pot is modeled by subdividing it into n contiguous reactors, n being greater than or equal to 2, and each of them is associated with sulfur oxide (SOx) masses and different temperatures; said step f) is modified in the following manner: f) in a first step, a rich time (tr) / lean time (tp) pair is chosen such that the downstream reactor of said pot, that is to say the last reactor considering the direction of movement of the gases, is freed of its surface oxygen (OS); f ') in a second step, a new rich time (tr) / lean time (tp) pair is chosen such that the said downstream reactor is completely free of its surface oxygen (OS); the determination of the rich time (tr) is carried out by at least one oxygen probe placed downstream of said pot or using a pattern of consumption of surface oxygen; said step f) is modified in the following manner: f) firstly, a rich time (tr) / poor time (tp) pair is chosen such that said rich time (tr) is long enough to get rid of together said pot of its surface oxygen (OS); f ') in a second step, a new rich-time (tr) / lean-time (tp) pair is chosen such that at least a portion of the reactors, namely at most the n-1 reactors situated downstream of the first reactor when considering the direction of movement of the gases, can not absorb again surface oxygen (OS). Other features and advantages of the invention will become apparent from the following detailed reading of a preferred embodiment. This description will be made with reference to the accompanying drawings in which: - Figure 1 is a diagram showing a device for carrying out the method according to the invention; FIG. 2A is a diagram showing a trap (catalyst) subdivided into four reactors, aligned between its inlet and its outlet; FIG. 2B is a diagram showing the level of OS within the trap, in an initial state prior to desulfation; FIG. 2C is a diagram showing the distribution of the sulfur oxides within the trap, in this same initial state; FIGS. 3A, 3C, 3D and 4A, 4C, 4D and 5A, 5C, 5D are diagrams corresponding respectively to those of FIGS. 2A, 2B and 2D, illustrating several successive stages of implementation of a form of FIG. embodiment of the invention; FIGS. 3B, 4B and 5B are diagrams showing the parallel evolution of the richness in the trap, during said successive steps.

The gasoline or diesel engine also comprises a plurality of fuel injectors 6 arranged so as to inject fuel into the combustion chamber that can operate the engine lean (normal operation). The computer 2 controls the operation of the injector 6 and the valve 5 to adjust the composition of the air / fuel supply of the engine, so that the richness of this mixture is momentarily greater than 1 (purge operation of the pot 4) in the combustion chamber. The computer 2 may optionally control an additional injector 7 of fuel placed in the exhaust line 3. A sensor 8 delivered to the computer 2 a signal representative of the temperature Tcat of the catalytic converter 4 or the temperature of the gases entering the pot. Other sensors (not shown) provide the computer with signals representative of other quantities conventionally used to manage the operation of the engine: engine speed, air intake pressure, etc. The catalytic converter 4 is naturally of the type comprising adsorption means (trap) of nitrogen oxides and sulfur when the engine operates lean air / fuel mixture. The computer 2 is moreover duly programmed in particular to: determine the current value, at a time of sampling i, of the mass of sulfur oxides absorbed by the pot 4, using a model; decide to trigger and stop a purge according to the mass of sulfur oxides stored in the pot 4 and the temperature thereof, this with the aid of an expert system, 30 - then order establishing the richness of the air / fuel mixture at a value corresponding to a stoichiometric or rich mixture, when such a purge is triggered, possibly ordering an increase in the temperature of the exhaust gas. The following will be described the structure and operation of the means necessary for the execution of the different functions of the system. 10 Operation in storage

This operation is observed when the engine is fed lean mixture. The calculation of the sulfur flow from the engine is based on knowledge of the sulfur content of the fuel and the oil required for engine lubrication. Since the sulfur contents are known, the estimation of the fuel flow rates and the oil consumption rate make it possible to evaluate the quantity of sulfur contained in the exhaust gas leaving the engine. From this data, it is estimated how much sulfur can be trapped in the NOx trap (NOxTrap). For this, a loading efficiency is defined which serves to characterize the SOx loading of the trap. Thus, according to the aforementioned EP-0 962 639: An instantaneous storage efficiency of SOx (EFFSOx) is defined as follows: EFFSOX = 1 - SOx flow at the exit of the exhaust line 3 SOX flow at the output of the motor 1 Moreover, a SSISSC filling ratio is defined by: SS / SSC = stored mass of SOx (SS) maximum mass of stored SOx (SSC) with SS = SOx Storage = mSOx, that is to say the sulfur mass trapped in the NOxTRAP at time t. SSC = SOx Storage Capacity, that is the SOx storage capacity, NOxTRAP-specific value, identified and mapped, varying with the temperature of the catalyst Tcat. According to the aforementioned patent, the storage efficiency EFFSOx is a function of the following parameters: the temperature of the catalyst Tcat, the ratio of the mass of SOx stored SS on the mass of maximum SOx storable SSC and the richness of the fuel Ri. In addition to the aforementioned parameters, it is possible to take into account the residence time of the gases or VVH (hourly volume velocity) as a function of the gas flow rate. This refines the model by this consideration. We have in the end:

EFFSOx = f (T, SSISSC, Ri, WH)

One can also take into account the thermal aging of the monolith in the estimation of the sulfur loading efficiency.

EFFSOx becomes then:

EFFSOx = f (T, SS / SSC, Wealth, WH, Aging) Finally, in loading mode, we have at the moment i: SS = mSOx; +1 = mSOx; + QSOx.EFFSOx

This also makes it possible to estimate the sulfur flow rate downstream of the NOxTRAP, in the loading mode: This flow rate is expressed as follows: QSOxAVAL = (1-EFFSOx) * QSOx Operation in storage of SOx The present invention is based on the hypothesis implying that the sulfur desorption takes place in SO2, H2S and OS form in rich medium (Ri> 1) once the surface oxygen of the trap (OS) is completely consumed by the reducing agents present in excess in the air / fuel mixture.

1st embodiment When purging SOx, it is sought to eliminate sulfur in a particular form: SO2, odorless species. But in some cases (wealth too high, hourly speed too low, ...), the sulfur can be emitted downstream of the pot 4, in the form of H2S and COS which are two smelly and toxic species at high concentration. The challenge of controlling desulfation is therefore to optimize the rate of desulphatation and to maximize the desorption of sulfur in the form of SO2 in the air.

In the context of the present application, it is assumed that the following reactions are obtained during desulphatation: • In rich mixture and in the presence of OS: reducing agents + OS • No desulfation but adsorption. • In rich mixture and in the absence of OS: reducers without OS ù H2S + COS + SO2, with possible conversion of SO2 in H2S + COS (this reaction is all the stronger as the time spent in the catalyst by the Desorbed SO2 is important). This means that when the conditions of temperature (about 700 ° C) and mixing (richness R;> 1) are reached, then: • As long as the OS is present in the trap, there is no effective desulfation . • Once there is no more surface oxygen in a part of the NOx Trap, the reductants will then react with the sulfur stored in the trap and cause the desorption of sulfur in the form of H2S, COS and SO2. following a kinetic law of desorption according to different parameters (for example richness, temperature, VVH ...) • In the presence of surface oxygen, the sulfur desorbed (SO2, H2S, COS) is re-adsorbed on the trap in SO2 form , following a kinetic law identical to that of the adsorption of sulfur in a poor medium, that is to say function of different parameters (for example richness, temperature, VVH, sulfur mass already stored ...). It is therefore sought to control the behavior of NOxTRAP in desulfation, taking into account the reactions (adsorption and desorption) which occur in the following cases: adsorption in the form of SO2, if the medium is generally poor, or rich in the presence Oxygen-desorption in the form of SO2, H2S, COS, if the medium is rich without surface oxygen. In this case, it can also be considered that there are two desorption kinetics, the first being the desorption in the form of SO2, the second being the desorption in the form of H2S and COS when the surface oxygen is consumed. The conversion of SO2 to H2S and COS in the presence of catalyst in a generally rich medium is also represented by kinetics, but this only characterizes the conversion rate and not the desulfation rate. Second embodiment

A sulfur loading / unloading model of a longitudinally discretized NOxTRAP in several contiguous reactors is considered here. Thanks to this configuration, it is possible to change the times of the rich time / lean time pair, for rich and poor settings on a given operating point, in order to optimize desulfation in the form of SO2 rather than in the form of H2S and COS. For this, the strategy here consists in decomposing the desulfation process in two successive periods: In a first step, the pair (rich time / lean time) is chosen so that the downstream part of the trap (for example the last reactor ) is never emptied of its surface oxygen. Thus, this reactor traps all the sulfur desorbed upstream, the interest being to trap the H2S and COS products from desorption or from the conversion of SO2 desorbed in the presence of reducing agents and catalysts. This first phase is maintained until the sulfur trapped in the upstream reactors has reached a predefined threshold or the sulfur charging in the last reactor has reached a predefined threshold.

• When the first phase is complete, the torque (rich time / lean time) is then chosen so that the downstream part of the trap (for example the last reactor) is completely emptied of its surface oxygen.

Thus, all the sulfur present in this reactor is predominantly desorbed as SO2 and does not have time to convert to H2S and COS (residence time too low).

3rd embodiment As for the previous embodiment, considering a sulfur loading / unloading model of a NOxTRAP discretized longitudinally in several reactors, this variant makes it possible to change the torque (rich time / poor time), for rich and poor settings on a given operating point, to optimize the rate of desulfation. For this, the strategy here consists in decomposing the desulfation process into two successive periods: • Firstly, the duration of the rich time is maintained so that the surface oxygen of the entire NOxTRAP is consumed. This duration can be known by the use of an oxygen sensor disposed downstream or by the use of a model for estimating surface oxygen. • In a second step, the couple (rich time / poor time) is chosen so that a downstream part of the trap (for example the downstream half) can never recover its surface oxygen. Thus, during the rich phases, the sulfur desorbed upstream of this part can not be re-adsorbed due to lack of oxygen. This strategy is particularly interesting for optimizing the rate of desulfation because it is mainly the upstream part of the NOxTRAP which is poisoned with sulfur. This limits the time required for downstream migration and has better desorption kinetics in the presence of oxygen upstream. With reference to FIGS. 2A to 5D, the following will be described below. of the second embodiment, as set forth above. The control of the fractionation and the phenomena that it induces impose to work with a discretized modelization or a certain number of sensors making it possible to know the conditions reigning in the various places of the NOx trap (in its length). In the following example of operation, we will base on a pot 4 (NOx trap) discretized into 4 reactors R1 to R4 arranged from upstream downstream when we consider the direction of passage of the gas through the pot. This is shown in Figure 2A. At the initial time of desulfation, the OS is homogeneously distributed in the trap, as shown in Figure 2B. Conversely, the distribution of SOx is unequal in length. The order of magnitude is as follows: 90% of the mass of SOx is contained in the first RI reactor NOx trap and 10% on the second R2. So there is no SOx in the last half of the trap (this is the situation of Figure 2C). In a first step, the desulfation starts with a rich mixture (Ri> 1).

The gearboxes that arrive at the NOx trap inlet start by reducing the OS present in the first reactor Ri. In the following three reactors R2 to R4, since the excess of reducing agents is entirely consumed by the first reactor, the mixture is stoichiometric: this is the situation of FIG. 3B. There are no more reductants present in the gases and there is therefore no reaction with the OS (FIG. 3C) in these reactors. As has been seen previously from the reactions involved, as long as there is OS in a reactor, there is no possible desorption of sulfur. Thus, the distribution of sulfur remains identical to the initial state during this step (FIG. 3D). In a second step, once the OS of the first reactor RI has been consumed in its entirety, the reducers will react with the sulfur and desorb it in the form of SO2, COS and H2S (with SO2 in a strong majority proportion), according to the law. kinetics of desorption.

At the same time, the OS of the second reactor R2 will decrease since there remains an excess of reducing agents at the outlet of the first reactor R1 (FIG. 4C).

Indeed, only a part of the excess of reductants will react with the SOx (unlike oxygen which completely oxidizes them in each reactor). There then remain reducers in the gases at the second reactor R2 which will reduce the OS thereof.

The sulfur species desorbed from the first reactor RI are carried away by the exhaust gases and are redeposited necessarily in the form of SO2 as soon as they are in a generally poor environment, following the kinetic law of adsorption. This is the case of the following reactor. Indeed, as long as the OS remains and the wealth Ri is equal to 1, because of the oxygen stored in the form of OS, it is then in a generally poor environment. Thus, there is "redeposition" of sulfur in SO2 form in the next (essentially) containing OS reactor. In the example, the sulfur contained by the first reactor R1 is moved in the second R2 (angled arrow of Figure 4D). In a third step and once the OS of the second reactor R2 has been completely consumed, the sulfur contained therein will then be desorbed, in the same way as for the first reactor. In the illustration of FIGS. 5A to 5D, since there remains OS in the third reactor R3, the sulfur desorbed in the first and second reactors will be deposited in the third essentially, and so on. Finally, before the OS of the NOx trap has been consumed (for example, the last non-regenerated reactor), it then goes into "poor mode". In the end, the sulfur that was contained in the upstream part of the NOx trap was moved to the most downstream part of it. A long "rich phase" period is then imposed to empty all the sulfur present in the last reactor of the trap. Thus, during the last desorption, the sulfur-containing products will spend only a short time in the wealth trap greater than 1. There is no longer a risk of conversion of SO2 into H2S-COS.

Claims (9)

1. Sulfur oxide (SOx) purge control method of a catalytic converter (4) for treating the exhaust gases of an internal combustion engine (1), said pot (4) comprising means for adsorption of sulfur oxides (SOx) contained in these gases, which comprises the following steps: a) the current mass (SSi) of sulfur oxides adsorbed in the pot (4) is evaluated at a sampling instant i, using the model: SSi + 1 = SSi + Qsox. EFFSOX where Qsox is the flow of sulfur oxides in the engine exhaust (1) and EFFSOX is the instantaneous storage efficiency of the sulfur oxides in the pot (4), b) a purge of the pot (4) is triggered ) when said quantity crosses a predetermined threshold (Si), a function of the temperature (Tcat) of the pot (4), by an increase in the richness (Ri) of an air / fuel supply mixture of the engine, starting from a richness corresponding to a lean or stoichiometric mixture, c) an alternation of injections of a rich mixture (Ri> 1) and then poor (Ri <_1), for so-called "rich time" respectively ( tr) and "lean time" (tp) characterized by the fact d) in step b), for said pot, the variation of the mass of sulfur oxides (SOx) purged per unit of time dmSOx / dt, in particular according to the temperature of the pot (Tcat), the richness (Ri) 25 of the air / fuel mixture, as well as according to a "stored mO2" quantity which Considering the quantity or the presence of surface oxygen (OS), this "surface oxygen" (OS) being the oxygen impregnated on the surface of the adsorption means of said pot, e) the desorption is considered sulfur only occurs when said richness (Ri) is greater than 1 and in the absence of surface oxygen (OS) f) the rich time (tr) / lean time (tp) pair of such way that the rich time (tr) is greater than the time required to consume the surface oxygen (OS) present in the pot (4). 2. Method according to claim 1, characterized in that the determination of said rich time (tr) is performed using means for detecting the presence or absence of surface oxygen (OS). 3. Method according to claim 2, characterized in that said detection means consist of at least one oxygen sensor. 4. Method according to claim 2, characterized in that said detection means consist of a pattern of consumption of surface oxygen (OS). 5. Method according to claim 1, characterized in that the determination of said rich time (tr) is greater than the consumption time of the surface oxygen (OS) and corresponds to the time for which a predetermined threshold of desorbed sulfur is reached . 6. Process according to Claim 1, characterized in that the said pot (4) is modeled by subdividing it into n contiguous reactors, n being greater than or equal to 2 and each of them being associated with masses of water. oxides of sulfur (SOx) and different temperatures. 7. Method according to claim 6, characterized in that said step f) is modified as follows: f) in a first step, a rich time (tr) / lean time (tp) pair is chosen such that the reactor downstream of said pot (4), that is to say the last reactor considering the direction of movement of the gases, is freed of its surface oxygen (OS). f ') in a second step, a new rich time (tr) / lean time (tp) pair is chosen such that the said downstream reactor is entirely free of its surface oxygen (OS) 8. Method according to claim 7, characterized in that the determination of the rich time (tr) is performed by at least one oxygen sensor placed downstream of said pot (4) or using a model of consumption of surface oxygen. 9. A method according to claim 6, characterized in that said step f) is modified as follows: f) in a first step, a rich time (tr) / poor time (tp) pair is chosen such that said time rich (tr) is long enough to rid the assembly of said pot (4) of its surface oxygen (OS); f ') in a second step, a new rich time (tr) / lean time (tp) pair is chosen such that at least a part of the reactors, namely at most the n-1 reactors situated downstream of the first reactor when we consider the direction of movement of the gases, can not absorb again surface oxygen (OS).