WO2007034095A1

WO2007034095A1 - Method for processing a measurement signal representing oxygen richness of a gas and corresponding device

Info

Publication number: WO2007034095A1
Application number: PCT/FR2006/050855
Authority: WO
Inventors: Jamil Abida; Aymeric Walrave
Original assignee: Renault S.A.S
Priority date: 2005-09-21
Filing date: 2006-09-07
Publication date: 2007-03-29
Also published as: FR2891010B1; FR2891010A1

Abstract

The invention concerns a method for processing a measurement signal representing oxygen richness of a gas, in particular a controlled ignition and injection motor vehicle exhaust gas, the measurement signal being established by means of a sensor filtering the said using at least one filtering element provided with a filtering coefficient. The method includes a step (51, 52, 53, 54, 55) of inversely transforming the measurement signal to as to determine a source signal (Vpredi) independent of the operating characteristics of the sensor, in particular the characteristics related to the filtering elements.

Description

Method for processing a measurement signal representative of the oxygen content of a gas and corresponding device

The present invention generally relates to the processing of a measurement signal derived from a sensor, in particular for regulating the richness of an air / fuel mixture admitted to a combustion engine of a motor vehicle with ignition and electronically controlled injection.

The invention is advantageously applicable to "lambda all-or-nothing" type probes.

The lambda probe is a sensor sensitive to the relative levels of oxidants and reducing agents in a gas. When placed in the exhaust duct of a heat engine, it provides information on the richness of an air / fuel mixture introduced into the combustion chamber. The output voltage of the lambda probe can switch between a high level (rich mixture) and a low level (lean mixture) located on both sides of a threshold corresponding to the stoichiometric ratio (richness "1"). For example, currently, it is considered that the lambda type probe measures a fuel-poor mixture (ie, a richness of less than 1) when it delivers a voltage below 450 mV and vice versa.

The output signal of the lambda probe is shaped in the electronic injection computer and the resulting logic information is a rectangular signal to which, by convention, the value "+1" is assigned when it is high and the value "- 1" when it is low.

To meet international standards for emissions of polluting substances, such as nitrogen oxides (NOx), unburned hydrocarbons (HC) and carbon oxides (CO), it is useful to improve the quality of the regulation of stoichiometric richness.

To do this, the lambda probe is commonly used, particularly in the context of a closed - loop control of the injection, when the exhaust system of the engine is equipped with a catalyst intended to reduce the emissions of these polluting components. exhaust gas. Indeed, this regulation makes it possible to enslave the richness of the air / fuel mixture admitted into the engine around the value "1", which is a prerequisite for the satisfactory combustion of the toxic components by the catalyst.

Therefore, it is important that the regulation of the mixture is accurate and has the lowest possible inertia especially in the first moments following the heating of the lambda type probe. Indeed, the lambda type probes need to be heated by a heat source other than the sole energy of the exhaust gas in order to deliver a usable signal, which correctly indicates the state of the mixture injected into the combustion chamber. However, in the first moments after the start of heating the probe, it can take a much longer time than hot to provide a voltage correlated with the true richness of the air / fuel mixture introduced into the combustion chamber. .

Therefore, when the probe has not sufficiently heated, it is observed when occurs a mixture rich transition / lean mixture, a period where the regulation is carried out for a rich mixture while the mixture has already become poor. Indeed, the probe gives information of a fuel-rich mixture a few more moments while the mixture observed has already become poor in fuel. This transition delay phenomenon disappears once the probe has sufficiently warmed up.

Currently, the problem is circumvented by delaying the wealth regulation loopback to ensure that the probe is ready to give a signal representative of the fuel richness of the gas but also that its reaction dynamics is good enough to avoid any inaccurate information.

However, the major disadvantage of this method is that during the period of time when the wealth is not regulated, it is sufficiently different from the optimal value for the vehicle to emit pollutants in large quantities.

Various methods propose to use the signal delivered by the probe to regulate the air / fuel mixture, such as EP-0-236 207 (Renault). However, this method does not relate to the processing of the signal delivered by the probe before its operation.

UFLO ("Ultra Fast Light Off") faster heating probes make it possible to reduce this delay without, however, eliminating it.

Consequently, the current state of the art does not make it possible to overcome the problem of the delay existing between the measurement of the richness of the gas and the variations of the effective wealth of the gas, and therefore of the poor regulation of the generated wealth. when the probe has not yet reached a sufficiently high temperature.

The invention aims to provide a solution to this problem. The object of the invention is to compensate for a physical phenomenon internal to the probe so as to eliminate the delay existing between the real value of the fuel richness of the gas and the value measured by the probe. For this purpose, the invention proposes a method of processing a measurement signal representative of the fuel richness of a gas, in particular the exhaust gas of a controlled ignition and injection vehicle, the signal of measurement being developed by means of a sensor filtering said gas using at least one filter element assigned a filter coefficient.

According to a general characteristic of the invention, the method comprises a step of inverse transformation of the measurement signal so as to determine a source signal independent of the operating characteristics of the sensor, in particular the characteristics related to the filtering elements.

In other words, the method of processing the signal delivered by the sensor, for example a lambda type probe, consists in summarily reversing the predominant and troublesome physical phenomena of the sensor, occurring during the heating phase of the sensor.

It is then possible to go back to the source information from the filtered information, which has the advantage of obtaining transitions of the measurement signal correctly synchronized with the actual transitions from passage of a fuel-poor gas to a gas rich in fuel or vice versa.

According to one embodiment of the invention, the inverse transformation step comprises:

a step of determining at least one order derivative

N of the measurement signal, - the calculation for each derivative of a coefficient depending on the filtering coefficients of the filtering elements, the calculation of said source signal from the measurement signal, the derivatives and the calculated coefficients. For example, said source signal is determined from the following relationship: _r . dy _r . d ² y _r _ d "yx = y + K, - - + K ₇ f + .... + ΛT - -, ^{y ι} dt ² dt ² " dt "where: x is the source signal, y is the signal measuring,

Ki, K ₂ , ..., K _n are coefficients dependent on the filter coefficients.

In addition, the method may advantageously comprise a step of filtering the derivatives of the measurement signal, prior to calculating said source signal.

The filtering can be for example of order 1.

According to one embodiment, it is possible to determine a first-order derivative of the measurement signal. According to one mode of implementation, the signals used may be voltages.

In a variant, the signals used may be values of the oxygen content in the gases.

According to one embodiment, the coefficient can advantageously be initialized to a value taken when the sensor gives an information distinct from its polarization value, and in which the coefficient is a function of time.

According to another aspect of the invention, there is provided a device for processing a measurement signal representative of the fuel richness of a gas, in particular the exhaust gas of a combustion engine for a motor vehicle with ignition and injection controlled by a control means, the device comprising a sensor mounted on the exhaust duct of the engine. Said sensor comprises at least one filter element having a filter coefficient and capable of filtering the gas so as to produce the measurement signal. According to a general characteristic of this other aspect of the invention, the control means comprises transforming means capable of performing an inverse transformation of the measurement signal so as to determine a source signal independent of the operating characteristics of the sensor, in particular characteristics related to the filter elements.

According to one embodiment, the transformation means comprises a shunt block able to determine at least one derivative of order N of said measurement signal, a calculation block connected to the shunt block and able to calculate for each derivative, a coefficient depending on the filtering coefficients of the filter elements of the sensor, and a block for generating the source signal as a function of the derivatives of the coefficients and of the measurement signal.

Preferably, said transformation means is able to determine the source signal according to the following relation:

" _{Dy τr} d ² y _τr d" y

¹ dt dt ² "dt" where: x is the source signal, y is the measurement signal,

K ₁ , K ₂ , ..., K _n are coefficients dependent on the filter coefficients.

According to one embodiment, the transformation means also advantageously comprises a filtering block connected between the shunt block and the source signal processing block, and able to filter the derivatives of the measurement signal.

For example, the filter block is capable of performing a filtering of order 1. According to one embodiment, the shunt block is able to determine a derivative of order 1 of the measurement signal.

According to one embodiment, the transformation means is able to receive and deliver signals in the form of voltages. In a variant, the transformation means is able to receive and deliver signals in the form of values of the oxygen content in the gases.

Preferably, said coefficient can be initialized to a value taken when the sensor gives information distinct from its polarization value, and in which the coefficient is a function of time.

The device as described above can be used in a system for regulating the richness of an air / fuel mixture admitted into a motor vehicle heat engine.

Other advantages and features of the invention will appear on examining the detailed description of an embodiment of the invention, in no way limiting, and the attached drawings, in which: FIG. 1 represents an embodiment of the invention; FIG. 2 shows the composition of an example of a probe used by the device according to the invention, FIG. 3 represents more precisely a module of the device according to the invention, FIG. implementation of the method according to the invention.

Referring to Figure 1. Reference 1 designates a spark ignition internal combustion engine.

The engine 1 is equipped with an injector 2 on the intake duct 3. The injector 2 receives a control signal from a microphone programmed computer 4 via a connection 5. The microcomputer 4 determines the nominal opening time of the injector 2 as a function of the air pressure measured by a pressure sensor 6 placed on the intake duct 3 and connected to the microcomputer 4 by a connection 7. In addition, the programmed microcomputer 4 determines the nominal opening time of the injector 2 from the speed of rotation of the motor V measured by a sensor 8a and delivered to the microcomputer 4 through an 8b connection.

The microcomputer 4 can also correct the nominal time as a function of other information such as, for example, the temperature of the atmospheric air T _alr delivered to the microcomputer 4 via a connection 9 or the temperature of the cooling water of the engine T _water delivered. to the microcomputer 4 by a connection 10.

To correct the nominal opening time of the injector 2, the microcomputer 4 also receives, via a connection 11, information delivered by a probe 12 disposed on the exhaust duct 13 of the engine 1. The signal the output of the probe 12 is shaped in the microcomputer 4. This signal contains information on the residual oxygen content of the exhaust gas and also on the momentary ratio of fuel and air of the mixture sucked by the engine 1 The high and low levels of this signal correspond to higher and lower riches respectively than the stoichiometric ratio (richness 1).

Referring now to Figure 2 which more precisely illustrates the constitution of a probe and its operation.

The gas whose wealth is to be measured is physically filtered by different elements of the probe, filtering at the end of which the wealth information is converted into a signal, here a voltage Vs. The existence of several successive filter elements shifts the moment when wealth information is effectively converted into voltage Vs.

The first filter element of the probe 12 is the protective cover of the probe 20. The latter allows the gases to pass only through an orifice 21. The second filter element is a porous protective material 22 in which the gases circulate to reach the sensitive element 23. This sensitive element 23 is the last filter element on which the oxygen molecules are fixed.

Each of these elements represents a filter applied to gases whose richness is evaluated, and thus implies a delay in the information to be measured. Each filter has a filter coefficient, inversely proportional to the importance of filtering, and which increases with temperature.

Then, the oxygen content detected on the sensitive element 23 is converted into a voltage Vs in comparison with the reference air

24. The voltage Vs is generated by the circulation of the O2 - ions in the electrolyte 23, and is measured by the electrodes 25 and 26.

Referring now to Figure 3 which more precisely illustrates the microcomputer 4. The microcomputer 4 receives as input the voltage Vs delivered by the probe with the aid of the connection 11, and the pressure P _alr , the temperature of the air T _alr , the temperature of the cooling water of the engine T _water and the speed V of rotation of the engine respectively by the connections 7, 9, 10 and 8b. From these different input information, the microcomputer 4 will then go back to the source information. To do this, the microcomputer 4 comprises in particular an inverse transformation block 30 which receives the voltage input Vs. More precisely, the voltage Vs is delivered to a first shunt block 31 whose role is to derive the input signal at least once. The shunt block 31 is connected by a connection 32 to a calculation block 33 able to calculate for each derivative delivered by the block 31, a coefficient which is a function of the filter coefficients of the different filtering elements of the probe 12.

The shunt block 31 is also connected to a filter block 34 by a connection 35.

Blocks 33 and 34 are connected to a multiplier 36 respectively by connections 37 and 38.

The multiplier 36 applies the different coefficients determined by the calculation block 33 to the corresponding derivatives filtered by the block 34. The multiplier 36 delivers the filtered and multiplied derivatives by a coefficient to an adder 39 via a connection 40. the adder 39 adds to the signal delivered by the multiplier, the voltage Vs delivered to the adder 39 by a connection 41.

The signal delivered by the adder 39 is delivered via a connection 42 to a block 43 which determines the nominal opening time of the injector.

Block 43 determines the nominal opening time of the injector of the vehicle also as a function of a signal delivered by a block 45 by means of a connection 44, and representative of the parameters

Pair, air, T _water and V. The microcomputer 4 then delivers via the connection 5 a control signal of the opening of the injector 2.

The flowchart shown in FIG. 4 illustrates an exemplary method implemented by the device represented in FIG. 3.

During a first step 50, a signal, for example a voltage, representative of the richness of the gas is measured by means of the probe. filtered. Alternatively, the measured signal could be the value of the oxygen level in the gas.

Then, the signal measured at time i, Usonde ;, is derived during a step 51. Indeed, since the probe performs a filtering of the source signal, it is possible to trace this source information by reversing the filtering performed. by the probe. In the case of a filter of order 1, we have: dy ldt = (x - y) - K, where: y is the filtered signal, x is the source signal, and K is the filtering coefficient.

The source signal x is then obtained by inverting the previous equation: x = y + ^ -IK. dt

Theoretically, for a filter of order N, we can find the source signal according to the same principle, from these different derivatives:

where the coefficients K ₁ , K ₂ ,..., K _n can be determined from combinations of the filter coefficients of each of the filter elements of the probe.

It is very likely that the physical filters of the probe are not perfect order 1 filters. However, the multiple derivatives and the conversions of tension / richness are heavy in software memory, in computation time and in loss of precision. In addition, the setting of the coefficients for each of the derivatives would be difficult to achieve. For these reasons, the study carried out here concerns an inversion of filters of order 1.

In addition, this simplification makes it much easier to manage the evolution of the coefficients K as a function of the temperature of the probe without degrading the desired effect, that is to say the detection of the rich / poor transition moment. .

It then comes, for the voltage Vs delivered by the probe:

Derivative, = (Vs ₁ - Vs _1-1 ) I TO, where: Derivative! is the derivative of the voltage Vs at instant i,

TO is the sampling step when acquiring measurements, i is the moment of measurement,

Vs is the measured voltage delivered by the probe. In step 51, the derivative of the signal measured according to the formula implementing only the first order derivative of the measured signal is thus calculated, so as to be able to reconstruct the source signal.

Once the derivative has been calculated (for a filter of order 1), the latter is filtered during a step 52. Indeed, because the probe voltage is a variable internal to a software, its derivative can not be that a succession of diracs, that is to say a very high value during a unit of time. Therefore, to compensate for this problem, apply a filter to the derivative of the probe voltage. This results in a derivative that is less chaotic and closer to the derivative of a continuous signal. However, since this filter causes a delay in probe information, it is important not to overfilter. It is therefore necessary to compromise between a noise-free signal and a dynamic response. For this, the filtering done is a first-order filtering in the application, whose discrete transfer equation is:

le- ^T0lτ I- _z ^'

H (z) = ^{-T0lτ z-} ¹ TO

or :

H (z) is the transfer function in discrete, TO is the sampling step (for example 8 mS), τ is the time constant of the filter. Filtering can be done by software. The formula used in this software is:

where: y is the filtered derivative, i is the instant considered, x is the probe voltage, and K is the filtering coefficient. Therefore, it comes:

TO -≡ τ = - and K = the ^τ .

Ln (I-K) The filtering of the derivative applied to the derivative of the voltage Vs obtained during the preceding steps gives, at a time i:

TO

Dér filtered _ = (Derived i - Dérjîltrée _{^ 1)} (1-e ^T) + Dér filtered _, _ _γ where:.

Filtered_drawn _t is the derivative filtered at time i, Derivative _t is the derivative at time i,

Filtered_drawn ^ ₁ is the filtered derivative at time i-1. During a step 53, a coefficient K is calculated for each derivative (only the first derivative for a filter of order 1). The coefficient K is initialized to a value when the probe is considered ready, that is to say ie when it provides a voltage sufficiently different from its bias voltage to consider that it gives a valid wealth information. This moment gives a relatively reliable indication of its temperature. Then we decrement the value of the coefficient K, over time. The decrementation method adopted is the following one, in order to have a big influence of the derivative at the start of the strategy, this influence decreasing rapidly over time: k = AI (B + time), where: the value of B allows to choose the rate of decrementation of the coefficient, such that more B is high, the more the decrease is slow.

The A / B pair allows you to choose the starting value of the coefficient.

It is recalled that the coefficient K is determined from a combination of the filtering coefficients of each of the filter elements of the probe.

Then, during a step 54, the coefficient K is applied to the derivative Der-filtered _t .

The product obtained and the probe voltage measured at time i are then combined in a step 55 according to the formula below if the derivative is negative:

V _predι = Vs, + Der_filtered, .K,

, and according to the formula below if the derivative is positive: V _predι = Vs, + Der _ filtered, .K.Coef _ positive, where Vp _r edi is the source voltage at instant i, from which the voltage measured by the probe results.

The parameter Coef_positif makes it possible to differentiate the passages poor / rich and the passages rich / poor. Indeed, since the filterings are not physically identical between rich / poor and poor / rich passage of a mixture, the coefficient Coef_positif is added to the coefficient applied to the derivative, when this one is positive. This coefficient also makes it possible, if necessary, to deactivate the strategy for transitions from a lean mixture to a rich mixture, where no prediction is really necessary for certain types of probe.

At the end of step 55, the voltage V _pre di is obtained and the coefficient K applied to the derivative as explained above is decremented. Consequently, by working on the source signal, the information delivered by the probe is exploitable much earlier, which makes it possible to determine the value of the wealth without waiting for the probe to reach a certain temperature. This further reduces pollutant emissions. Moreover, it is possible to deactivate the method described above when the probe has reached a sufficient temperature, and no longer introduces a delay in the information it delivers.

Of course, the invention is in no way limited to this specific example and can be applied to any type of spark ignition engine, regardless of the number of injectors and cylinders that equip it. Likewise, the parameters for calculating the nominal opening time of the injection are given solely by way of examples and it is possible, among other things, to use an air flow sensor instead of a pressure sensor. mounted on the intake duct. Moreover, if the signal coming from the probe is not too parasitized, or when the software used is fine enough, it is possible to eliminate the filtering step of the derivative.

It is also possible to calculate a derivative of an order greater than 1 in order to approximate the theoretical study.

In the example described above, the voltage is converted into richness at the end of the process. However, it is possible to convert the voltage from the probe as soon as it is measured and to work throughout the process with the value representative of the richness.

Claims

Process for the treatment of a measurement signal representative of the fuel richness of a gas, in particular the exhaust gas of a controlled ignition and injection vehicle, the measuring signal being produced by means of a sensor filtering said gas with at least one filter element having a filter coefficient, characterized in that the method comprises a step (51, 52, 53, 54, 55) of inverse transformation of the measurement so as to determine a source signal (V _pre di) independent of the operating characteristics of the sensor, in particular the characteristics related to the filter elements.

The method of claim 1 wherein the reverse transformation step comprises:

a step of determining (51) at least one derivative of order n of the measurement signal,

the calculation (53) for each derivative of a coefficient (K) depending on the filtering coefficients of the filter elements,

calculating (55) said source signal from the measurement signal, the derivatives and the calculated coefficients.

The method of claim 2, wherein said source signal is determined from the following relationship:

^{y ι} dt ² dt ^{2 n} df where: x is the source signal, y is the measurement signal,

K1, K2, ... Kn are coefficients dependent on the filter coefficients.

4. Method according to the preceding claim, further comprising a filtering step (52) of the derivatives of the measurement signal, prior to calculating said source signal.

5. Method according to the preceding claim, wherein the filtering is of order 1.

6. Method according to one of claims 2 to 5, wherein a derivative of order 1 of the measurement signal is determined.

7. Method according to one of the preceding claims wherein the signals used are voltages.

8. Method according to one of the preceding claims wherein the signals used are values of the oxygen content in the gases.

9. Method according to one of claims 2 to 8, wherein said coefficient (K) is initialized to a value taken when the sensor gives an information distinct from its polarization value, and wherein the coefficient is a function of time.

10. Apparatus for processing a measurement signal representative of the fuel richness of a gas, in particular the exhaust gas of a combustion engine for a motor vehicle with ignition and injection controlled by a control means (4), the device comprising a sensor (12) mounted on the exhaust duct (13) of the engine (1), said sensor (12) comprising at least one filter element having a filter coefficient and capable of filtering the gas in a manner developing the measurement signal, characterized in that the control means (4) comprises transformation means (30) adapted to perform an inverse transformation of the measurement signal so as to determine a source signal independent of the operating characteristics of the sensor, in particular the characteristics related to the filter elements.

1 1. Device according to claim 10, wherein the transformation means comprises a shunt block (31) able to determine at least one derivative of order n of said measurement signal, a calculation block (33) connected to the block of derivation and capable of calculating for each derivative a coefficient dependent on the filtering coefficients of the filtering elements of the sensor, and a generating block (39) of the source signal as a function of the derivatives, the coefficients and the measurement signal.

12. Device according to one of claims 10 or 1 1, wherein said transformation means is adapted to determine the source signal according to the following relationship:

"Dy" d ² y " ^dn y ^yl dt ² dt ² " df where: x is the source signal, y is the measurement signal,

Kl, K2, ... Kn are the coefficients dependent on the filter coefficients.

13. Device according to the preceding claim, wherein the transformation means further comprises a filter block (34) connected between the branch block and the source signal development block, and able to filter the derivatives of the measurement signal. .

14. Device according to the preceding claim, wherein the filter block is capable of performing a filter of order 1.

15. Device according to one of claims 1 1 to 14, wherein the shunt block is adapted to determine a derivative of order 1 of the measurement signal.

16. Device according to one of claims 10 to 15, wherein the transformation means is adapted to receive and deliver signals in the form of voltages.

17. Device according to one of claims 10 to 15 wherein the converting means is adapted to receive and deliver signals in the form of values of the oxygen content in the gases.

18. Device according to one of claims 1 1 to 17 wherein said coefficient is initialized to a value taken when the sensor gives an information distinct from its polarization value, and wherein the coefficient is a function of time.

19. Use of a device according to one of claims 10 to 18 in a system for regulating the richness of an air / fuel mixture admitted into a motor vehicle heat engine.