FR2746851A1 - Closed-loop control of mixture of fuel and air for combustion engine - Google Patents

Abstract

The method is applied to an engine (1) having a catalytic convertor (7) with upstream and downstream oximeters (5,6). In one loop a correction ( alpha CL) is calculated for the basic fuel injection time (Ti) as a function of the signal from the upstream oximeter (5). In the other loop at least one of the parameters of the first loop is adapted as a function of the signal (SO2DW) from the downstream oximeter (6). At least one parameter of the second loop is adapted to the integration time of the convertor, defined as the delay in propagation of an upstream mixture variation to the downstream side.

Description

 The present invention relates to a method and a device for closed-loop regulation of the richness of an air / fuel mixture intended for supplying an internal combustion engine, and more particularly to such a method and device taking into account counts the ability of a catalytic converter to purify engine exhaust.

 The prior art, in particular from patent US Pat. No. 4,750,328, is known for a system for regulating the air / fuel mixture in which a first regulation loop determines a correction of a basic injection time as a function of a signal delivered by an oxygen sensor placed upstream of a catalytic exhaust. As shown in FIG. 1, when the signal delivered by the probe switches from a level indicating a rich mixture to a level indicating a lean mixture (or respectively from poor to rich), a value is added (resp. Subtracted) predetermined P at the current correction, and this correction is incremented (resp. decremented), at a predetermined frequency or recurrence, with a second value I as long as the signal from the probe remains at this level. A predetermined delay TD can be applied between the tilting of the signal from the probe and the addition of the value P. The values P, I, and TD thus form the parameters of the first regulation sulk which aims to regulate the richness of the mixture around an average richness depending on these parameters. A second regulation boude comes to adapt at least one of the parameters of this first boude according to a signal delivered by an oxygen sensor placed downstream of the pot to adjust the average richness of the mixture to a value, for example to the stoichiometry (A = 1), predetermined by a setpoint to which the signal from the downstream probe is compared. This second sulk is executed with a predetermined frequency, lower than that of the first sulk. Such a system has certain drawbacks. It has for example been observed that when the average richness fixed by the first loop becomes greater than the set point of the second loop, there is a significant delay for the signal from the downstream probe to react to this change, delay conditioned by the capacity of the catalytic converter to absorb fluctuations in wealth. As a result, the adaptation of the parameters of the first boude continues during this delay, for example in the direction of an enrichment of the mixture, even when the latter is already too rich or vice versa, which leads to exceeding the Absorption capacity of the pot and generate pollution.

 The present invention therefore aims to provide a richness control method which overcomes these drawbacks and takes into account the state of the catalytic converter in the process.

 These objects of the invention are achieved, as well as others which will appear in the following of this description, by means of a closed-loop regulation method of the richness of an air / fuel mixture intended for feeding. of an internal combustion engine according to which a correction of a basic fuel injection time is calculated, in a first regulation loop, as a function of a signal supplied by a first oxygen sensor disposed upstream of '' a catalytic converter placed in the engine exhaust gas stream, and, in a second regulation loop, at least one of the parameters of this first sulk is adapted as a function of a signal supplied by a second probe located downstream of the pot, characterized in that at least one of the parameters of the second sulk is adapted as a function of an integration time of the catalytic converter, this time being defined as the time necessary for a variation in richness to be present at Am have the pot is found downstream of it.

 According to an important characteristic of the present invention, the integration time of the catalytic converter is estimated as a function of the gas flow rate on the one hand, and of the oxygen storage capacity of the catalytic converter on the other hand.

 According to a first embodiment of the invention, the considered parameter of the second loop is a sampling period and this period is an inverse function of the gas flow rate. Advantageously, this is determined from the switching period of the signal from the upstream probe.

 According to a second embodiment of the invention, the parameter considered of the second sulk is at least one of the gains of the sulk and the oxygen storage capacity of the pot is determined as a function of a parameter representative of the signal dynamics of the downstream probe.

 According to an important characteristic of this embodiment, the parameter representative of the dynamic range of the signal from the downstream probe is calculated by integrating the absolute value of the difference between the raw signal and the filtered signal from the downstream probe. Advantageously, this parameter is also a function of the gas flow rate passing through the catalytic converter.

 According to another important characteristic of this embodiment, the gains of the second sulk are canceled when the filtered signal of the downstream probe is included in a predetermined area on either side of a reference value.

 The invention also relates to a device specific to the implementation of the regulation method defined above. This device is distinguished in that the second regulation means use adjustable parameters and in that it comprises filtering means suitable for filtering the signal delivered by the downstream probe and adaptation means for calculating from the signal the downstream probe, of the filtered signal and of the signal delivered by the flow measurement means, a signal representative of an integration time of the catalytic converter, and to deduce therefrom the parameters of the second regulation means.

Other characteristics and advantages of the invention will appear on reading the description which follows and on examining the appended drawings in which:
Figure 1 shows the time graphs of the signal from the upstream probe and the
correction made by the first sulk, as described in the preamble,
FIG. 2 schematically represents an engine equipped with a suitable device
for the implementation of the invention
FIG. 3 represents the integration time of the catalytic converter as a function on the one hand
the flow of gas passing through it and secondly its state of wear,
FIG. 4 represents the filtered signal of the downstream probe as a function of the richness of the
gas leaving the pot,
FIG. 5 represents the time graphs of the wealth at the outlet of the pot
catalytic, of the signal of the downstream probe, raw and filtered, and of the parameter representative of
the dynamics of this, respectively for a new and used pot, and
Figures 6 and 7 show the method according to the invention in the form of
flowchart.

 We now refer to Figure 3, to explain the concept of integration time of a noted catalytic converter, as it will be used later. This integration time is defined as the time required for a variation in richness present upstream of the pot to be found downstream of the latter. This figure shows the variation in the integration time of a catalytic converter as a function of the gas flow Q passing through it on the one hand and on the other hand as a function of the age or the degree of wear of the pot. It can thus be seen that the integration time decreases with wear, which is explained by the reduction in the capacity of the materials used to store the oxygen present in exhaust gases poor in unburnt hydrocarbons or to restore it. to rich gases. Likewise, this time decreases when the gas flow rate increases. It can be noted that the two factors on which the integration time depends have different rates of evolution. For example, the wear of the catalytic converter is a slow phenomenon, which extends over several months, even several years under normal conditions of use, while the influence of the flow changes rapidly, of the order of a few milliseconds to a few seconds. It therefore appears that a method for regulating the richness of the air I fuel mixture which uses richness information taken downstream of the catalytic converter to adjust the average richness value around which is centered a regulating valve located upstream of that -this must take into account a delay which depends on this integration time. In fact, when this integration time is maximum, as in the case of a new pot combined with a low gas flow rate, for example under engine idle speed, a drift in the mean value of the upstream sulk in the direction excess fuel will only be detected after this integration time. Corrective action on the parameters of the upstream regulation loop will then take place late, and, the drift having continued in the same direction during this time, pollution of unburnt hydrocarbons will follow until the effect of this correction is felt. Of course, a reverse drift, caused for example by this correction, will produce analogous effects, with other pollutants generated by a too lean mixture.

 We will now describe, in relation to FIG. 2, a device suitable for implementing a method of regulating the richness which takes into account the influence of this integration time of the catalytic converter. This figure shows an internal combustion engine 1 controlled by an electronic control computer 2. This engine is equipped with a sensor 3 located in the intake duct to measure the mass flow of gas Q passing through the engine, d '' a sensor 4 of the engine rotation speed N, and a catalytic converter 7 on either side of which are located two oxygen probes 5 and 6. These probes respectively provide signals representative of the oxygen richness of the gases exhaust, S02UP upstream of pot 7 and S02DW downstream of it. These various signals are supplied to the computer 2 which in turn generates a control signal Ti of at least one fuel injector 8 located on the engine.

Conventionally, a block 21 determines a basic injection time Tib from the flow rate Q and speed N information. To obtain the control signal Ti from the injector 8, this basic injection time Tib is corrected in a correction block 22 by a correction coefficient pc, developed by a first regulation block 23. Of course, other corrections, not shown, can also be applied in block 22, such as for example corrections as a function of the coolant or air temperature entering the engine. This QCL correction coefficient is developed in the regulation block 23 from the richness signal SO2UP delivered by b probe 5 according to the process already mentioned in the preamble in relation to FIG. 1 and which will be detailed below in relation with FIG. 6. According to this method, at least one of the parameters P, I and / or TD used in block 23 is modified to slave the average richness to a richness reference value S02DWOBJ.

In the following description, we will consider a modification of the parameters of block 23 relating to a TROFF offset added to the predetermined value of parameter P, by way of example only, the modification possibly also relating to the other parameters, individually or in combination.

 According to the invention, the richness signal S02DW delivered by the probe 6 is filtered by a filter 25. It can be seen by referring to FIG. 4, that the use of a filtered signal has certain advantages. In this figure, the richness of oxygen in the exhaust gases in the vicinity of the probe 6 is shown on the abscissa, and the value of the probe signal on the ordinate. It thus appears that the filtered signal S02TRMY has a less steep slope than the raw signal S02DW, particularly in the vicinity of B = 1, which makes it possible to obtain better regulation stability. The filtered signal S02TRMY is compared with the setpoint SO2DWOBJ. The deviation (S02TRMY S02DWOBJ) is supplied to a second regulation block 24 which modifies the parameters of block 23.

 As we saw earlier, it is important to take into account the integration time of the catalytic converter in determining this offset. In a first embodiment of the invention, we are more particularly interested in the influence of the flow factor. In this case, block 24 calculates cyclically, with a sampling period Tch, the offset TROFF by adding to the offset calculated at the preceding cyde a fixed value of the same sign as the deviation (S02TRMY - S02DWOBJ). This sampling period is determined according to an inverse function of the gas flow rate. In this way, the higher the gas flow, the shorter the period and therefore the faster the TROFF offset changes, thus allowing faster correction. Conversely, at a low flow rate, the offset evolves slowly, thus eliminating too frequent oscillations of the correction buoy managed by block 24. Advantageously, it has been observed that the period with which the signal SO2UP from the upstream probe 5 switches from a lean to rich level or vice versa is an inverse function of the gas flow rate passing through the engine. Thus, by synchronizing the calculation of the TROFF offset in block 24 with the transitions of the upstream probe, a sampling period T.d is obtained, which is an inverse function of the gas flow rate without the need to perform any calculation.

 Advantageously, one can define around A = 1 a richness range kp.

in which the efficiency of the catalytic converter is optimal, which results in a band of values of S02TRMY materialized by A1 and A2 around the set value, for which the difference is canceled (SO2TRMY - SO2DWOBJ). This has the effect of blocking the evolution of the TROFF offset and therefore making it possible to stay longer in this optimal richness range.

We will now explain a second embodiment of the invention in which the parameters of the regulating block 24 are also adjusted as a function of the wear of the catalytic converter, estimated from its oxygen storage capacity. Reference is made to FIG. 5 in which time graphs have been represented of the oxygen content X of the exhaust gases leaving the catalytic converter, of the signal from the downstream probe, raw and filtered, and of a parameter representative of the dynamics of this, and this for the same fluctuation in the richness of the gases entering the pot. The left part of the figure corresponds to a new pot, and the right to a used pot. It can be seen on the oxygen content curve that it is substantially stable in the case of a new pot, whereas for a used pot, the fluctuation present upstream is found downstream, consequence of the loss of the capacity of the pot to absorb and restore oxygen, due to the aging of the materials that compose it. Consequently, for a new pot, the signal S02DW of the downstream probe 6 is practically confused with the same filtered signal S02TRMY whereas it is clearly distinguished from it for a used pot. This observation is advantageously used to calculate a parameter
DWMOVE representative of the signal dynamics of probe 6 by integrating the absolute value of the difference between the raw signal (S02DW) and the filtered signal (S02TRMY) of the downstream probe (hatched part). This parameter, represented in dotted lines on the bottom curve, will serve as an estimator of the oxygen storage capacity and therefore of the TS integration time of the pot.

In this case, the regulation performed by block 24 is of the proportional and integral type, i.e. the TROFF offset is calculated according to the formula
TROFF = TROFFP + TROFFI in which the proportional part
TROFFP = Kp * (S02TRMY - S02DWOBJ) and the integral part TROFFI at time i is equal to TROFFI *,) + Ki * (S02TRMY - SO2DWOBJ) where Kp and Ki are gains. An adaptation block 26 is then provided to adjust the gains from this regulation as a function of the DWMOVE parameter representative of the signal dynamics of the downstream probe. Block 26 receives for this purpose the value of the raw signal S02DW of the probe 6 and that of the filtered signal
S02TRMY from filter 25. It can also receive the gas flow rate information Q and reuse it in the calculation of the DWMOVE parameter in order to make it dependent on the flow rate. Block 26 then calculates the DWMOVE parameter using the formula: DWMOVE = DWMOVE + Km Q (lSO2DW - SO2TRMYl - DWMOVE) where Km is a predetermined filtering coefficient. Depending on this parameter, the adjustment block 26 reads from stored Tabi tables (DWMOVE) and
Tabp (DWMOVE), obtained experimentally during the development of the engine, the gains Ki and Kp which will be used in block 24 for determining the offset
TROFF. Advantageously, the block 26 checks whether the filtered signal S02TRMY belongs to the band of values determined by Ar and 2 around the set value S02DWOBJ, and in this case imposes Ki = O and Kp = O, which has the effect of blocking the evolution of the TROFF offset when the richness is within an optimal efficiency range of the catalytic converter.

 We will now describe, in relation to the flow diagrams of FIGS. 6 and 7, an example of the method according to the invention, in which use has been made of a combination of the two modes of implementation previously described. The flowchart in Figure 6 details a program for calculating the coefficient vola. This program is executed by the computer 2 with a fixed recurrence, for example every 8 milliseconds. In step 101, the computer reads the respective values of the signals from probes 5 and 6 (S02UP and SO2DW) as well as the value of the flow rate Q of gas absorbed by the engine from the sensor 3. In step 102, the filter 25 and adjustment block 26 respectively calculate the values of S02TRMY and DWMOVE. Note that these two values are calculated each time a measurement is made. We go to step 103 where we test whether the signal from the upstream probe 5 is indicative of a rich mixture. If so, the state of an indicator Fg which has the value 1 is checked in step 110 if the mixture was already rich during the previous execution of the process.

In this case, the OCCL coefficient is decremented by a value I in step 111 and the process ends at 130. If the indicator Fg is equal to 0, there is a transition of the signal from the probe 5 from poor to rich. We update the status of the indicator
Fg in step 112 and go to step 113 to calculate the TROFF offset. This calculation is carried out by the subroutine illustrated in FIG. 7. It will be noted that this subroutine is only executed during a transition of the probe 5, in accordance with the first mode of implementation described. In this subroutine, it is tested, in step 151, if the filtered signal SO2TRMY of the downstream probe 6 is included in a predetermined interval around the setpoint SO2DWOBJ. If this is the case, we go to step 152 in which the gains Kp and Ki are fixed respectively at 0 and 1. Otherwise, we go to step 153 where these gains are extracted from the Tabi tables ( DWMOVE) and
Tabp (DWMOVE). In all cases, we then proceed to step 154 where the TROFF offset is calculated as a function of the determined gains and of the difference between the setpoint SO2DWOBJ and the filtered value SO2TRMY. We end in 155 with a return instruction which returns to step 114 of FIG. 6, where the coefficient aCt is calculated by subtracting the predetermined value P, modified by the offset TROFF, from the previous value of ri. If, in step 103, the test of the signal from the upstream probe 5 reveals a lean mixture, steps 120 to 124 are carried out in a similar manner to steps 110 to 114. It will again be noted that at step 123, the TROFF offset calculation subprogram of FIG. 7 is only executed during a transition, this time from rich to poor, of the signal from the upstream probe 5.

 Of course, the invention is not limited to the embodiment described above which has been given only by way of example. As already mentioned, the modification of the parameters of the first boude can also relate to the integration value I or to the delay TD. or any combination of these parameters. Similarly, the value of the gas flow rate Q may not be measured but estimated according to other measurements such as for example the pressure prevailing in the intake manifold or the opening of a throttle valve and the engine speed.

Claims (10)

1. A method of regulating closed mud of the richness of an air / fuel mixture intended for the supply of an internal combustion engine (1), according to which a correction (< xcL) of a basic fuel injection time (Ti), as a function of a signal supplied by a first oxygen sensor (5) disposed upstream of a catalytic converter (7) placed. in the exhaust gas flow from the engine, and, in a second regulation valve, at least one of the parameters (P, l, T0) of this first valve is adapted as a function of a signal (S02DW) supplied by a second probe (6) located downstream of the pot, characterized in that at least one of the parameters (Kp, Ki, Tech) of the second boude is adapted as a function of an integration time (TE) of the pot catalytic, this time being defined as the time necessary for a variation of richness present upstream of the pot to be found downstream of the latter. 2. Method according to claim 1 characterized in that the integration time (TZ) of the catalytic converter is estimated as a function of the gas flow rate (Q) on the one hand, and of the oxygen storage capacity of the catalytic converter d 'somewhere else. 3. Method according to claim 2 characterized in that the considered parameter of the second mud is a sampling period (Tech) and in that this period is an inverse function of the gas flow rate. 4. Method according to claim 3, characterized in that the inverse function of the gas flow (Q) is determined from the switching period of the signal from the upstream probe 5. Method according to claim 2, characterized in that the parameter considered of the second sulk is at least one of the gains (Kp, Ki) of the sulk and in that the oxygen storage capacity of the pot is determined by function of a parameter (DWMOVE) representative of the signal dynamics of the downstream probe. 6. Method according to claim 5, characterized in that the parameter representative of the dynamic range of the signal from the downstream probe is calculated by integrating the absolute value of the difference between the raw signal (S02DW) and the filtered signal (S02TRMY) downstream probe. 7. Method according to claim 6, characterized in that the parameter representative of the dynamic range of the signal from the downstream probe is also a function of the gas flow rate (Q) passing through the catalytic converter. 8. Method according to claim 7, characterized in that the parameter representative of the dynamic range of the signal from the downstream probe is calculated by the formula: DWMOVE (i) = DWMOVE (i-1) + Km * Q (ISO2DW (i) - S02TRMY (i) l - DWMOVE (i-1)) where Km is a predetermined coefficient and Q represents the gas flow rate. 9. Method according to any one of claims 5 to 8, characterized in that the gains (Kp, Ki) of the second loop are canceled when the filtered signal (S02TRMY) of the downstream probe is included in a predetermined area of share and other of the reference value (S02DWOBJ>. 10. Device for implementing the method according to any one of claims 1 to 9, of the type comprising  means (3) for measuring a gas flow (Q) passing through the engine (1) and the pot  catalytic (7),  a first oxygen sensor (5) capable of determining the richness of the gases  exhaust upstream of the catalytic converter,  a second oxygen sensor (6) capable of determining the richness of the gases  exhaust downstream of the catalytic converter,  a computer (2) adapted to provide control signals of at least one  fuel injector (8), said signals being produced by first and second  regulation means (23; 24) as a function of the signals coming from the means of  measurement of gas flow and oxygen sensors and a richness value of  setpoint, characterized in that  the second regulation means (24) use adjustable parameters and in that it also includes  filtering means (25) adapted to filter the signal delivered by the downstream probe (6),  adaptation means (26) for calculating from the signal of the downstream probe  (S02DW), the filtered signal (S02TRMY) and the signal (Q) delivered by the  flow measurement a signal (DWMOVE) representative of an integration time (TZ) of the  catalytic converter, and deduce the parameters of the second regulation means.