FR3075269A1 - METHOD FOR DETECTING PRE-IGNITION OF FRESH AIR AND FUEL MIXTURE - Google Patents

Abstract

 The invention relates to a method for detecting a pre-ignition of a mixture of fresh air and fuel in a cylinder (4) of an internal combustion engine (1), said internal combustion engine being controlled to run at a set speed and operating according to successive combustion cycles.  According to the invention, when the engine is used in a stabilized operating mode, for example when the variations in the reference speed over three successive combustion cycles remain below a predetermined threshold, a pre-ignition of the mixture is diagnosed during the third of the three successive combustion cycles when two successive decreases in the average engine speed sufficiently large are detected during these three successive combustion cycles, each of these two decreases being accompanied by a significant decrease in the average combustion torque calculated by torque meter software, a first drop in the average torque between the first and second of the three cycles and a second drop in the average torque between the second and the third of the three cycles.

Description

The present invention relates generally to the reduction of polluting emissions from internal combustion engines.

It relates more particularly to a method of detecting a pre-ignition of a mixture of fresh air and fuel in a cylinder of an internal combustion engine, said internal combustion engine being controlled to run at a set speed and operating according to successive combustion cycles by turning at an average speed and developing an average torque which can be calculated by a software torque meter.

Technological background

We are currently looking, within an ever more restrictive legislative framework and with a concern for the preservation of the environment, for technical solutions making it possible to reduce the quantities of polluting molecules released by internal combustion engines into the atmosphere.

One solution is to market small displacement internal combustion engines.

But we also want to find solutions that preserve the performance of these engines (in terms of power and torque available).

These small displacement engines are then designed to operate at higher pressure levels, that is to say that they are adapted to compress the mixture of fresh air and fuel present in the cylinders at higher pressures. than in larger displacement engines.

These small displacement engines are therefore likely to be used at critical operating points, in particular at high load and at low speed.

Indeed, on these operating points, there may appear a phenomenon known as pre-ignition which greatly reduces the performance of the engines and which is also potentially destructive for these engines.

This phenomenon corresponds, in a spark-ignition engine, to an autonomous ignition of the mixture of fresh air and fuel, before its triggering by a spark plug.

This ignition of the mixture in the cylinder before the piston has reached its top dead center is then liable to oppose the movement of the piston and therefore to slow down the rotational movement of the engine crankshaft.

Therefore, we want to be able to quickly and precisely detect the appearance of this pre-ignition phenomenon, so that we can quickly change the engine settings to make it disappear.

The methods of detecting this phenomenon which are currently known and implemented are not entirely satisfactory, insofar as this phenomenon is not detected with sufficient reliability, in particular at low speed at which it is difficult to find a compromise satisfactory between non-detection and false pre-ignition detection.

In fact, at low speed, the behavior of the engine is sensitive to resonances coming from the transmission chain (gearbox, transmission shafts, wheels, tire contact on the road). These resonances tend to distort the values of the coefficients used in the estimation of the torque by software torque meter, such as for example the torque meter which is used in patent document FR 17 52730 not yet published on the date of filing of this application.

Object of the invention

In order to remedy this drawback, the present invention provides a reliable and inexpensive detection solution, in the sense that it does not require for its implementation any component which is not already installed in the engines of vehicles present on the market.

For its implementation, it uses a software torque estimation method of the average torque produced during a combustion cycle of the gas mixture in an engine cylinder.

By software torque meter is meant in a manner known per se a method in which said average combustion torque is calculated from the running times, in front of a running scroll sensor fixedly mounted in the vicinity of a crown secured to a flywheel of the engine or of an engine crankshaft, of a plurality of measurement marks arranged on said crown.

Such a software torque meter is known in particular from numerous patents, for example from the publication FR-A1-2681425 to which reference may be made. We can also refer, without limitation, to other publications such as FR-A1-2793558, EP-A1-0448603, etc.

More particularly, a detection method as defined in the introduction is proposed according to the invention, in which, when the variations in the reference speed over three successive combustion cycles remain below a predetermined threshold, a pre-ignition is diagnosed of the mixture during the third of the three successive combustion cycles when, during these three successive combustion cycles, are detected:

a first drop in average speed, between the first and the second cycle, which is greater than a first predetermined speed threshold, and a first drop in the value of average torque estimated by said torque meter, between the first and the second cycle, said first decrease in the average torque value being greater than a first predetermined torque threshold; then,

a second drop in average speed, between the second and the third cycle, which is greater than a second predetermined speed threshold, and a second drop in the value of average torque estimated by said torque meter, between the second and the third cycle, said second drop in the average torque value being greater than a second predetermined torque threshold.

In this first diagnostic phase, the detection of a drop in torque between the first and second cycles makes it possible to be more precise in the detection of a pre-ignition at low speed than the overestimation by the software torque meter alone. value of the torque of the second cycle which is provided for in the patent application not yet published FR 17 5270. In fact, according to the method according to the invention, the two couples are estimated by the same software torque meter and therefore suffer substantially the same bias estimation due to imprecision at low speed. The method which is the subject of this invention therefore makes it possible to demonstrate a reduction in real torque, independently of the estimation method, by canceling the bias linked to the method.

As will be detailed in the remainder of this description with reference to the figures, the Applicant has found that the phenomenon of pre-ignition is accompanied by a drop in torque on the one hand between the first and the second cycle and on the other hand part between the second and the third cycle. It is therefore on the basis of this observation that the invention proposes to rapidly diagnose this pre-ignition phenomenon.

As each of the consecutive couples compared is estimated in the same way by the software torque meter, they undergo substantially the same estimation bias, which improves the detection accuracy of the pre-ignition phenomenon at low speed.

The process which is the subject of the invention is then inexpensive to implement in the sense that only the torque and the speed of the engine must be estimated in order to be able to make a diagnosis. However, these parameters are generally already estimated on current vehicles, for the implementation of other functionalities.

Other non-limiting and advantageous characteristics of the detection method according to the invention, taken individually or in any technically possible combination, are the following:

a step of estimating a first value of the average speed during the first of the three successive combustion cycles is provided, a step of estimating a second value of the average speed during the second of the three combustion cycles successive, and a step of comparing a difference between the first value and the second value of the average speed with the first speed threshold, then a pre-ignition of the mixture is diagnosed during the third of the three successive combustion cycles provided in particular that said difference reveals a first drop in average speed, of amplitude greater than said first speed threshold;

- there is provided a step of estimating a first value of the average torque during the first of the three successive combustion cycles, a step of estimating a second value of the average torque during the second of the three combustion cycles successive, and a step of comparing the difference between the first value and the second value of the average torque with the first torque threshold, then a pre-ignition of the mixture is diagnosed during the third of the three successive combustion cycles provided in particular that said deviation is greater than said first torque threshold;

- there is provided a step of estimating the second value of the average torque during the second of the three successive combustion cycles, a step of estimating a third value of the average torque during the third of the three successive combustion cycles , and a step of comparing the difference between the second value and the third value of the average torque with the second torque threshold, then a pre-ignition of the mixture is diagnosed during the third of the three successive combustion cycles provided in particular that said deviation is greater than said second torque threshold; and

a step of estimating the second value of the average speed during the second of the three successive combustion cycles is provided, a step of estimating a third value of the average speed during the third of the three successive combustion cycles , and a step of comparing the difference between the second value and the third value of the average speed with the second speed threshold, then a pre-ignition of the mixture is diagnosed during the third of the three successive combustion cycles provided in particular that said difference is greater than said second speed threshold.

The invention also relates to a method for controlling an internal combustion engine, comprising:

- a step of developing a set of instructions, including a deposit regime,

a step for controlling the internal combustion engine according to said instructions,

a diagnostic step during which a detection method is implemented, then, if a pre-ignition is detected,

- a step of correcting at least one of said instructions.

The invention also relates to an internal combustion engine comprising:

- at least one cylinder inside which a piston slides,

- a crankshaft which is connected to the piston,

- a means of estimating the average speed, suitable for estimating the speed of rotation of the crankshaft,

a means of estimating the average torque, adapted to estimate the torque exerted by the piston on the crankshaft, and

- a computer adapted to implement a detection method as mentioned above.

Detailed description of an exemplary embodiment

The description which follows with reference to the appended drawings, given by way of nonlimiting examples, will make it clear what the invention consists of and how it can be carried out.

In the accompanying drawings:

- Figure 1 is a schematic sectional view of an internal combustion engine according to the invention;

- Figure 2 is a graph illustrating the variations of the instantaneous speed and the average speed of the internal combustion engine of Figure 1, in the absence of pre-ignition phenomenon;

- Figure 3 is a graph illustrating the variations of the instantaneous speed and the average speed of the internal combustion engine of Figure 1, in the presence of pre-ignition phenomenon; and

- Figure 4 is a flowchart illustrating the steps for implementing a detection method according to the invention.

In FIG. 1, an engine block 2 of an internal combustion engine 1 is shown diagrammatically.

This engine block 2 has four main parts, including a cylinder block 3, an oil sump 5 which is fixed under the cylinder block 3 and which contains oil intended to lubricate the various engine components, a cylinder head 6 which is fixed on the cylinder block 3 and a cylinder head cover 7 which covers the cylinder head 6.

The cylinder block 3 internally delimits at least one cylinder 4. In practice, it will be considered here that the cylinder block 3 delimits four cylinders 4 in line with vertical axes A10.

Conventionally, each cylinder 4 houses a piston 8 which is adapted to slide along its internal wall in an alternating rectilinear movement (or reciprocating movement) of axis coincident with the axis A10 of cylinder 4.

The piston 8 has a peripheral skirt which is pierced transversely with two openings for receiving an axis on which is engaged an upper end of a connecting rod 9. The lower end of this connecting rod 9 is linked, via 'an eccentric connection to a crankshaft 10. Thus, the reciprocating rectilinear movement of the piston 8 makes it possible to rotate the crankshaft 10 of the internal combustion engine 1 around its longitudinal axis, called the engine axis A1.

In the remainder of this presentation, the lever, the projection, in a plane orthogonal to the axis A10, between the motor axis A1 and the axis A2 of the eccentric will be called "lever arm". This lever arm is therefore proportional to a cosine function of the angle formed between the longitudinal axis of the connecting rod 9 and the axis A10.

For the admission of fresh air to each cylinder 4, the cylinder head 6 is pierced with an intake duct 11 which extends from an air distributor 12 to an intake opening 13 provided in a lower face of the cylinder head 6.

For the exhaust of burnt gases outside each cylinder 4, the cylinder head 6 is pierced with an exhaust duct 14 which originates in an exhaust opening 16 adjacent to the intake opening 13 and which opens into an exhaust manifold 15 fixed to the cylinder head 6.

Valves 17, 18 provided in these conduits are used to regulate the flow rates of fresh air inlet and burnt gas outlet in each cylinder 4.

There is further provided here, for each cylinder 4, an injector 20 making it possible to inject fuel into the cylinder 4. This injector 20 is fixed in the cylinder head 6 and opens into the cylinder 4.

The internal combustion engine 1 is here with positive ignition. It then comprises, for each cylinder 4, at least one spark plug 19 which is fixed in the cylinder head 6 and which opens into the cylinder 4. This spark plug 19 is designed to initiate the combustion of the mixture of fuel and fresh air in cylinder 4.

It should be noted in this regard that the internal combustion engine 1 operates according to the Beau de Rochas cycle, with an intake time, a compression time, a combustion-expansion time, and an exhaust time.

It is planned to start the combustion of the mixture in an offset manner in the four cylinders. Thus, when one of the cylinders is at the intake time, a second is at the compression time, a third at the combustion-expansion time, and a fourth at the exhaust time.

For simplicity, in this presentation, when we will deal with "successive combustion cycles /, i + 1, i + 2", we will consider in practice more precisely combustion-expansion times occurring successively in different cylinders of the engine. Otherwise formulated, the expression “successive combustion cycles /, i + 1, i + 2” is used for simplification, but the more exact expression would rather be “combustion-expansion time /, i + 1, i + 2 intervening successively in the engine cylinders ".

In support of FIG. 1, the internal combustion engine 1 also comprises a means for estimating an instantaneous speed of the crankshaft 10, from which an estimate of the average speed Qj of the engine can be deduced during each cycle of combustion of order i and an estimate of the average torque C ,, or average combustion torque developed during said cycle of order i.

The average speed Ω, corresponds to the average of the rotational speed of the crankshaft 10 during a combustion cycle (that is to say during a combustion-expansion time). It is therefore intended to vary in stages between the different successive combustion cycles i, i + 1, i + 2.

In the following description, we will make the difference between average speed Qj (on a cycle of order i), instantaneous speed R and setpoint speed Ω _ο , this last parameter corresponding to the speed of rotation at which we want the engine is running.

The average torque C, meanwhile corresponds to the average of the torque exerted by each piston 8 on the crankshaft 10 during a combustion cycle in the cylinder in which the piston 8 considered slides. It is therefore understood that, at each combustion cycle, only the contribution of the piston 8 considered is taken into account for the calculation of this average torque C ,. It is therefore intended to vary in stages between the different successive combustion cycles /, i + 1, i + 2.

As shown in FIG. 1, the means for estimating the instantaneous speed R here comprises, on the one hand, a crown which is integral in rotation with the crankshaft 10 and which carries measurement marks 30, 31 regularly distributed over its contour, and, on the other hand, a sensor (not shown) for detecting the measurement marks 30, 31 which is fixedly mounted in the vicinity of the crown.

The crown is shown here as being fixed to the crankshaft but it could alternatively be secured therein in rotation, for example via an engine flywheel.

Here it bears markers 30 which are all identical, with the exception of a wider marker 31 making it possible to readjust the measurements in the event of deviation from these measurements.

The sensor then makes it possible to detect the running time of each of the measurement marks 30, 31. The instantaneous speed of rotation R or instantaneous speed R is deduced from the time of running between two consecutive marks.

This sensor is connected to a computer which is adapted to determine the average speed or average speed of rotation Ω and the average combustion torque Ci taking into account the running time of each of the measurement marks 30, 31 in front of the sensor .

The determination of the mean torque Cj is carried out in the manner described in the publication FR-A1-2681425.

This estimated torque for the order i combustion cycle is calculated using the equation:

(Equ. 1) C, = -a Ω, 'Ej cos4> j + b Ω ² , equation in which:

- Ω, designates the average angular speed during the cycle,

- E, cos <t> j denotes a projection, on the reference line of references relating to the angular periods of combustion, of the alternative component E, of the instantaneous angular velocity of the measurement references placed on the crown integral with the flywheel engine inertia, at the combustion frequency, and

- a and b are constants determined experimentally.

The computer here is formed by the central computer of the vehicle, the one which controls the various organs of the internal combustion engine 1, in particular the fuel injector 20 and the spark plug 19.

This computer conventionally comprises a processor (CPU), a random access memory (RAM), a read only memory (ROM), various input and output interfaces.

Thanks to its input interfaces, the computer is adapted to receive input signals from various sensors, in particular from the above-mentioned scroll sensor.

The ROM stores data used in the process described below.

In particular, it stores a predetermined map on a test bench by which the computer is adapted to generate, for each operating condition of the engine, speed and torque thresholds (which will be well described in the rest of this presentation).

The read-only memory also stores a computer application, consisting of computer programs comprising instructions whose execution by the processor allows the method described below to be implemented by the computer.

This application also comprises, in a conventional manner, programs for calculating various parameters and various instructions, among which:

- the engine air load (which corresponds to the ratio of the actual engine air filling divided by the maximum engine filling at a given speed and at a given pressure),

- the speed setpoint Q _c at which it is desired that the crankshaft 10 rotate,

- an advance ignition instruction,

- a fuel injection instruction.

Finally, thanks to its output interfaces, the controller is suitable for transmitting output signals to the various components of the engine, in particular to the fuel injectors 20 and to the spark plugs 19.

Conventionally, when the driver of the motor vehicle switches on the ignition, the computer initializes and then controls a starter and the fuel injectors 20 so that the latter start the engine.

It calculates in parallel a set of setpoints, including a setpoint speed, a fuel injection setpoint, an ignition advance setpoint, etc. When the engine is started and it is controlled according to these instructions, the fresh air taken from the atmosphere is regularly blown into each cylinder 4 to mix with fuel. This so-called combustion mixture is then burned in the cylinders 4, thanks to the spark plugs 19 provided to initiate combustion.

Each spark plug 19 is thus controlled to initiate combustion at the end of the compression time, generally a little before the piston 8 reaches its highest point in the cylinder 4 (called top dead center).

The ignition advance corresponds to the difference separating the ignition of the spark plug from top dead center (it is expressed in degrees of rotation of the crankshaft 10).

It sometimes happens that the combustion mixture self-ignites before the spark plug 19 has been piloted to start the combustion of the mixture. We are talking about pre-ignition.

The present invention therefore proposes a solution for detecting this pre-ignition phenomenon.

This process is implemented only when certain conditions of stability of the engine operating point are met, and in particular when the variations in the reference speed Q _c over three successive combustion cycles /, i + 1, i + 2 remain below a predetermined SQ _C threshold.

Apart from these stability conditions, the process is not implemented since it would provide unreliable results and since the risks of pre-ignition are very reduced. When the engine accelerates, the ignition advance is such that there is generally no pre-ignition phenomenon. Conversely, when the engine decelerates, the quantity of fuel injected is so reduced that there is generally no more phenomenon of pre-ignition.

In general, the detection method according to the invention consists in considering three successive combustion cycles /, i + 1, i + 2 occurring in three separate cylinders, and in diagnosing a pre-ignition of the mixture during the third of these cycles if the computer detects two successive drops in average speed, the first drop being accompanied by a first drop in the average torque estimated by the software torque meter, and the second speed drop being accompanied by a second drop in the average torque estimated by the software torque meter . We can refer to Figures 2 and 3 to explain this phenomenon.

In FIG. 2, the variations of the instantaneous speed R have thus been represented in the absence of pre-ignition phenomenon.

The setpoint regime Q _c being considered constant, these variations are identical during each of the three successive combustion cycles i, i + 1, i + 2. We will then describe here only one of these combustion cycles (that is to say a single combustion-expansion time occurring in a cylinder 4, more precisely the cycle of order i + 1).

It can be seen in this FIG. 2 that, during a first time interval S1, the instantaneous speed R increases, which is due to an increase in the torque exerted by the piston considered on the crankshaft. This increase in torque is produced by the start of combustion of the mixture and by the increase in the lever arm.

There is then observed, over a second time interval S2, a stagnation of the instantaneous speed R. This stagnation is due to the fact that the combustion of the mixture in the cylinder continues stably and to the fact that the lever arm varies slightly.

There is then observed, over a third time interval S3, a reduction in the instantaneous speed R. This reduction is due to the fact that the lever arm is reduced and that the combustion is ending.

Finally, we observe, over a fourth time interval S4, which starts at the moment when the piston passes the top dead center, a continuation of the reduction of the instantaneous speed R.

In FIG. 3, the variations of the instantaneous speed R have thus been represented in the presence of a phenomenon of pre-ignition in the third cylinder (the one where the third combustion cycle / + 2 considered) takes place.

The first combustion cycle /, which takes place in a first of the cylinders 4, takes place in exactly the same way as on the cycles of order i, i + 1 or i + 2 represented in FIG. 2.

The second and third combustion cycles i + 1, i + 2, which take place in second and third cylinders 4, take place in a different way.

It is observed that, during the second combustion cycle i + 1, the first and second time intervals S1, S2 are identical to those illustrated in FIG. 2, for the same reasons.

We then observe, over a third time interval S3 ’, a decrease in the instantaneous speed R greater than that expected without pre-ignition. This sharp decrease at the end of this time interval S3 ′ is due to the fact that the pre-ignition of the mixture in the third cylinder generates a braking torque on the crankshaft 10. As a result, the average speed decreases between the cycle d order i and the cycle of order i + 1, whereas in the absence of pre-ignition, it should have remained substantially constant.

On a fourth time interval S4 ’, the engine time of the third cycle i + 1 begins with the combustion already in progress, and the pressure is high as soon as the top dead center is passed.

During a fifth time interval 85, the instantaneous speed R increases, but more slowly than in the absence of pre-ignition. This relative slowness is explained by the fact that, although the lever arm increases, the combustion of the mixture in the third cylinder is finite. The torque is lower than that which would have been expected in the absence of pre-ignition.

Finally, during the sixth time interval 86, there is a decrease in the instantaneous speed R. The low torque decreases until the end of the engine time of the order cycle i + 1.

Finally, the consequence of this instantaneous speed profile during the time intervals S4 ', S5 and S6, is a significant reduction in the estimate of the average torque between cycle i and cycle i + 1, as well as between cycle i + 1 and the cycle i + 2.

FIGS. 2 and 3 also show the variations in the average speed Ω over each of the three combustion cycles considered.

In Figure 2, we observe that the average speed Ω, or average speed of rotation Ω, remains constant from one combustion cycle to the next.

On the other hand, in FIG. 3, it is observed that the average speed Ω decreases a first time between the combustion cycle / and the combustion cycle i + 1, and a second time between the combustion cycle i + 1 and the cycle of combustion i + 2.

In Figure 4, there is shown a flowchart illustrating an example of algorithm allowing the computer to implement the invention.

This algorithm is intended to start, in step E1, at the start of each combustion cycle in a cylinder 4. The duration of implementation of this algorithm being of three successive combustion cycles, it is therefore understood that the algorithm will be at each instant implemented several times in parallel by the computer, in a sliding manner.

In step E2, the computer waits for the first two successive combustion cycles i, i + 1 to occur.

Then, the following step E3 makes it possible to check whether the operating conditions of the engine are such that it is possible to establish a diagnosis.

This step consists more precisely in verifying that at least one control parameter satisfies a predetermined stability criterion.

More precisely here, the computer verifies that the reference speed R _c , the engine air load and the ignition advance satisfy, on the first two successive combustion cycles i, i + 1, a predetermined stability criterion .

As an example, this criterion can be that the variation of each of these control parameters between the first two successive combustion cycles i, i + 1 is less than 5%.

If at least one of these control parameters does not satisfy this stability criterion, the algorithm is interrupted.

On the contrary, if these control parameters all satisfy the stability criterion, the algorithm continues in a step E4.

During this step E4, the computer acquires:

- a first value Ωι of the average speed Ω during the first combustion cycle /,

- a second value Ω _{ί + 1} of the average speed Ω during the second combustion cycle i + 1,

- a first value Ci of the average torque C estimated by software torque meter during the first combustion cycle /,

- a value Cj + i of the average torque C estimated by software torque meter during the second combustion cycle i + 1.

We use for example the equation 1 which we transpose to the cycle of rank i + 1:

Ci + i = -a Ωί ₊ ι 'E _{j +} i cos0 _{i +} i + b Ω, + ι ²

During the following step E5, the computer calculates the difference ΔΩ, + ι between the first value Ω, and the second value Ω _{ί +} ι of the average speed, according to the following mathematical formula:

ΔΩϊ + ι ⁼ Ω | - Ω _{ί + 1}

Then, it compares this difference ΔΩ, ₊₁ with a first speed threshold SC _{i + 1} .

If this difference ΔΩ, + ι is less than or equal to the first speed threshold SC _{i + 1} , which means that there has been no significant drop in average speed, the algorithm continues at step E12 .

This step E12 consists for the computer in deducing that there is no pre-ignition phenomenon in the third cylinder.

On the contrary, if this difference ΔΩ _{ί + 1} is greater than the first speed threshold SC _{i + 1} , which means that there has been a significant drop in average speed, the algorithm continues in a step E6.

During this step E6, the computer calculates the difference AC _{j +} i between the first value Ci and the second value Cj + i of the torque estimated by software torque meter, according to the following mathematical formula:

ACj + i - Cj - Cj + - |

It then compares this difference AC _{j +} i with a first torque threshold SC _{j +} i.

If this difference AC _{j +} i is less than or equal to the first torque threshold SC _{j +} i, which means that there has been no significant decrease in the estimated torque, the algorithm continues with step E12.

On the contrary, if this difference AC _{i +} i is greater than the first torque threshold

SCi + i, which means that there has been a notable decrease in the torque estimated by software torque meter, the algorithm continues with a step E7.

During this step E7, the computer waits for the third of the successive combustion cycles / + 2 to occur.

Then, during step E8, it checks whether the engine operating conditions are still such that it is possible to establish a diagnosis.

This step is identical to step E3 above, and will therefore not be described again.

Thus, if at least one of the control parameters does not satisfy the stability criterion, the algorithm is interrupted.

On the contrary, if these control parameters all satisfy the stability criterion, the algorithm continues with a step E9.

During this step E9, the computer estimates:

- a third value Ω _{ί + 2} of the average speed Ω during the third combustion cycle / + 2, and

- a third value C _{j + 2} of the torque estimated by software torque meter during the third combustion cycle i + 2.

We use for example the equation 1 which we transpose to the cycle of rank i + 2:

C _{i + 2} = -a Ω _{ί + 2} E _{i + 2} cos0 _{i + 2} + b Ω | ₊₂ ²

Then, the computer calculates the difference ΔΩ _{ί + 2} between the second value Ω, + ι and the third value Ω, ₊₂ of the average speed, according to the following mathematical formula:

ΔΩ | + ₂ = Ωί + ι - Ω | + ₂

It then compares this difference ΔΩ, ₊₂ with a second speed threshold 8Ωη- ₂ .

If this difference ΔΩ _{ί + 2} is less than or equal to the second speed threshold δΩ _{ί + 2} , which means that there has been no significant drop in average speed, the algorithm continues on step E12 .

On the contrary, if this difference ΔΩ _{ί + 2} is greater than or equal to the second speed threshold δΩί ₊₂ , which means that there has been a second marked drop in average speed, the algorithm continues at a step E10.

During this step E10, the computer calculates the difference ΔΟ, + ₂ between the second value C _{i +} i and the third value C _{i + 2} of the torque estimated by software torque meter, according to the following mathematical formula:

ACj + ₂ ^- Cj + 1 - Cî + 2

It then compares this difference ΔΟ, + 2 with a second torque threshold SCj + 2.

If this difference ΔΟ, + 2 is less than or equal to the second torque threshold SCj + 2, which means that there has been no significant decrease in the estimated torque, the algorithm continues with step E12.

On the contrary, if this deviation AC _{i + 2} is greater than the second torque threshold SCj + ₂ , which means that there has been a significant drop in the torque estimated by software torque meter, the algorithm continues with a step E11.

This step E11 consists for the computer in deducing that there has been a pre-ignition phenomenon in the third cylinder.

In this case, there is preferably a step for correcting the engine setpoints to prevent this phenomenon from happening again.

For the implementation of the algorithm shown in FIG. 4, therefore, several speed and torque thresholds SCj + i, SCj + 2, SQ _{j +} i, SQî + _{2 are used} .

These thresholds will all be read in the cartography recorded in the ROM. Note that they may vary depending on the engine operating conditions.

The present invention is not limited to the embodiment described and shown, but the skilled person will be able to make any variant according to the invention. It is possible in particular to use any other known method of determining the average combustion torque by software torque meter, in particular the methods described in the publications FR-A1-2793558, EP-A10448603, etc.

Claims (7)

1. Method for detecting a pre-ignition of a mixture of fresh air and fuel in a cylinder (4) of an internal combustion engine (1), said internal combustion engine (1) being controlled to turn to a set speed (Q _c ) and operating according to successive combustion cycles (i, i + 1, i + 2) by turning to an average speed (Ω) and developing a given average torque (C) of combustion , said mean speed (Ω) and mean torque (C) being calculated by a software torque meter from the running times, in front of a running motion sensor fixedly mounted in the vicinity of a crown secured to a crankshaft (10) of the internal combustion (1), of a plurality of measurement marks (30, 31) arranged on said ring, characterized in that, when the variations in the reference speed (Ω _ο ) over three successive combustion cycles (i, i +1, i + 2) remain below a predetermined threshold (δΩ ₀ ), pre-ignition is diagnosed of the mixture during the third of the three successive combustion cycles (i + 2) when, during these three successive combustion cycles (i, i + 1, i + 2) are detected: - a first average speed drop (Ω), between the first and the second cycle, which is greater than a first predetermined speed threshold (δΩ _{ί + 1} ), and a first drop in the estimated average torque value (C) by said torque meter, between the first and the second cycle, said first drop in the average torque value (C) being greater than a predetermined first torque threshold (SCj + 1); then, - a second drop in average speed (Ω), between the second and third cycles, which is greater than a second predetermined speed threshold (8Ωί ₊₂ ), and a second drop in the value of average torque (C) estimated by said torque meter, between the second and third cycles, said second drop in the mean torque value (C) being greater than a second predetermined torque threshold (SC _{i + 2} ). 2. Detection method according to claim 1, comprising: - a step of estimating a first value (Ω,) of the average speed (Ω) during the first of the three successive combustion cycles (i), a step of estimating a second value (Ω _{ί +} ι) of the average speed (Ω) during the second of the three successive combustion cycles (i + 1), and a step of comparing a difference between the first value (Ω,) and the second value (Ω, + ι) of the average speed (Ω) with the first speed threshold (δΩ _{ί +} ι), and in which one diagnoses a pre-ignition of the mixture during the third of the three successive combustion cycles (i + 2) provided in particular that said difference reveals a first drop in the average speed (Ω), of amplitude greater than said first speed threshold (8Ωί + ι). 3. Detection method according to the preceding claim, comprising: - a step of estimating a first value (Ci) of the average torque (C) during the first of the three successive combustion cycles (i), a step of estimating a second value (Cj + i) of the average torque (C) during the second of the three successive combustion cycles (i + 1), and a step of comparing the difference between the first value (Cj) and the second value (C _{i +} i) of the average torque (C) with the first torque threshold (SCi + i), and in which a pre diagnosis is made ignition of the mixture during the third of the three successive combustion cycles (i + 2) provided in particular that said difference is greater than said first torque threshold (SCj + i). 4. Detection method according to one of the preceding claims, comprising: a step of estimating the second value (C _{j +} i) of the average torque (C) during the second of the three successive combustion cycles (i + 1), - a step of estimating a third value (Cw) of the average torque (C) during the third of the three successive combustion cycles (i + 2), and a step of comparing the difference between the second value (C _{j +} i) and the third value (Cj + ₂ ) of the average torque (C) with the second torque threshold (SC _{i + 2} ), and in which we diagnose pre-ignition of the mixture during the third of the three successive combustion cycles (i + 2) provided in particular that said difference is greater than said second torque threshold (SC _{i + 2} ). 5. Detection method according to one of the preceding claims, comprising: a step of estimating the second value (Ω _{ί +} ι) of the average speed (Ω) during the second of the three successive combustion cycles (i + 1), a step of estimating a third value (Ω _{ί + 2} ) of the average speed (Ω) during the third of the three successive combustion cycles (i + 2), and a step of comparing the difference between the second value (Ω _{ί +} ι) and the third value (Ωί ₊₂ ) of the average speed (Ω) with the second speed threshold (δΩ _{ί + 2} ), and in which a pre-ignition of the mixture is diagnosed during the third of the three successive combustion cycles (i + 2) provided in particular that said difference is greater than said second speed threshold (8Ω _{ί + 2} ). 6. Method for controlling an internal combustion engine (1), comprising: - a step of developing a set of setpoints, including a setpoint regime (Ω _ο ), a step for controlling the internal combustion engine (1) according to said instructions, - a diagnostic step during which a detection method in accordance with one of the preceding claims is implemented, then, if a pre-ignition is detected, - a step of correcting at least one of said instructions. 7. Internal combustion engine (1) comprising: - at least one cylinder (4) inside which slides a piston (8), - a crankshaft (10) which is connected to the piston (8), a means for estimating the average speed (Ω), suitable for estimating the speed of rotation of the crankshaft (10), and a means of estimating the average torque (C), adapted to estimate the torque exerted by the piston (8) on the crankshaft (10), characterized in that it comprises a computer adapted to implement a detection method according to one of claims 1 to 6.