FR2997498A1 - Method for diagnosing exhaust gas recirculation system in internal combustion engine of car, involves developing control of recirculation valve to reduce error, and reporting system fault if error exceeds interval of tolerance - Google Patents

Abstract

The method involves assessing (102) an error (epsilon) by subtracting a set of state value (Qcons) from a state value (Qcal) calculated or measured from a pressure (P) inside an engine cylinder block when an internal combustion engine runs in higher richness or equal to a unit value. A control (Cr) of a recirculation valve is developed (105), so as to reduce the error, and an exhaust gas recirculation system fault is reported (106) if the error exceeds an interval of tolerance (epsilon min, epsilon max) where the calculated state value is equal to cumulated energy release by a driving cycle. Independent claims are also included for the following: (1) a computer program for diagnosing an exhaust gas recirculation system in an internal combustion engine (2) an exhaust gas recirculation system in an internal combustion engine.

Description

The invention relates to a method for the diagnosis of a gas recirculation system in an internal combustion engine and the gas recirculation system which implements a system for recirculating gas. This method of diagnosis, especially in a motor vehicle Gas recirculation systems and related diagnostic methods are already known in the prior art EP1076170 discloses a system for detecting failure of a recirculation device dc exhaust gas which is measured a gas pressure in the EOR recirculation line for an open position and a closed position of a recirculation valve, so as to detect a failure when the difference between the two measured pressures does not The problem here is to be able to establish a diagnosis whatever the degree of openness of Thus, US6763708 discloses a method for diagnosing gas flow restriction in an EGR recirculation system by comparing inlet tubing pressures with expected pressures. in inlet tubing for different engine speeds. The problem posed here is to be able to establish a diagnosis whatever the dispersions of manufacture and / or mounting of the engine. The object of the invention is to respond to the problems posed by the prior state of the art, in particular for a mode of carburation in richness greater than or equal to a unit value. In order to achieve this objective, the subject of the invention is a method of diagnosis of a gas recirculation system in an internal combustion engine. comprising steps of: evaluating an error by subtracting a state value setpoint from a comp value calculated or measured when the engine is running in richness, p (ricure or equal to a unit value; control (c recirculation valve so as to reduce said error; - report a defect in the gas recirculation system if said error goes out of a tolerance range. Advantageously, said state value is calculated or measured from Preferably, said state value is a cumulative energy release per engine cycle, and in particular, said set value value is stored in at least one engine cylinder. In particular, the method comprises a step of determining said consig- ration of the vehicle's recirculation system. The engine state is determined from a predetermined value map each for a rotation value and for an effective average engine pressure in greater value, equal to the unit value. The method is also advantageous when it comprises a step consisting in predetermining said state variable setpoint in the test phase so as to obtain a good acoustic control of the engine and / or in order to obtain a good thermal control of the reaction. The invention also relates to a computer program comprising program code instructions for executing the steps of the method according to the invention when said program is in accordance with the invention. The invention also relates to a gas recirculation system in an internal combustion engine comprising at least one recirculation valve, a motor cylinder pressure sensor and a calculator comprising: a computer program according to the invention.

In particular, said calculator comprises a state variable setpoint table, indexed by a motor rotation speed and / or by an effective pressure. Other features and advantages will become apparent on reading the description which follows, with reference to the appended drawings, in which FIG. 1 represents an internal combustion engine to which the invention is applicable, FIGS. To 2e are chronourams which show a way of modifying a richness of the supply of the engine, FIGS. 3a and 3b are graphs of the pressure in a cylinder during the last two engine cycle times, respectively in richness of less than and greater than 1, FIG. noise and temperature variation curves observed as a function of a burnt gas recirculation rate. FIG. 4 shows a change in energy release, as a function of a recirculation rate. FIGS. 6a and 6b show tables of values of co one respectively in high pressure recirculation and in low pressure recirculation, FIG. 7 is a simplified process logigrumme according to the invention. not. In a manner known per se with reference to FIG. 1, four cylinders, without this number being limiting, of an engine block 10, receive an oxidizing gas through an intake manifold 16. After combustion of a fuel The fresh air 11 is sucked in at low pressure (HP) from the outside by a compressor 14 of a tuloo compressor to feed the tubing. admission 16 in ga. high pressure oxidant (11,). The compressor 14 is driven by a turbine 18 of the turbo compressor, preferably of variable geometry to relax 1-s burnt gases from the exhaust pipe 17 which are routed us-t 'to a trap 19 to nitrogen oxides generally nominated NoxTrap. The flue gases are then. discharged at low pressure (BP) in a muffler 20 at the outlet of the trap 19. Still in known manner, the fresh air 11 is filtered at the inlet by an air filter I 2 disposed upstream of the compressor 14. A cooler 15 downstream of the compressor makes it possible to run against the temperature rise of the compressor by the compression. An admission flap 61 downstream of the cooler 15 is connected to a digital or wired bus 60 so as to control the closing opening by a computer 50 itself connected to the bus 60 to modulate the flow rate from introduced into the intake manifold. A flue gas recirculation valve 62, commonly known as EGRHP (Acronymec 1 e I English expression 1 :, x11a List Ga ', Recirculation Iiigh Pressure) makes it possible to take off at high pressure a portion of the flue gases in the exhaust manifold. 17 to reintroduce them into the intake manifold 16 so as to obtain an oxidizing gas which thus combines the reintroduced portion of flue gases with the air sucked from the outside. The manna EGPIlP 62 is connected to the bus 60 so as to command its opening and closure by the computer 50. A gas recirculation valve 64, commonly known as EGRLP (acronym for the English expression Fxhaust Gas Recirculation Low Pressure) makes it possible to remove at low pressure a portion of the burnt gases in the muffler 20 to reintroduce them upstream of the compressor 14 so as to alternately obtain an oxidizing gas which thus combines the reintroduced portion of flue gases with the air sucked from the 'outside. The EGRLP valve 64 is connected to the bus 60 in order to command 1-opening and the closure by the computer 50. A cooler 22 makes it possible to lower the temperature of the reintroduced flue gas so as to make it easier for the compressor to compress. 14.

An exhaust flap 63 in the muffler 20, is connected to the bus 60 so as to control the opening and closing by the computer 50 to modulate the flow of gas burned to the atmosphere in the form of gas 21 and thus facilitate sampling by the valve EGRLP A flowmeter disposed downstream of the air filter 12 comprises a sensor 51 coiii-iccté bus 60 to indicate the flow of air Q sucked to the calculator SU, pleur of temperature 52 , disposed in the intake pipe 16, is connected to the bus 60 to indicate to the computer 50 a combustion gas temperature T52 called. The oxygen probe 53, located downstream of the turbine 18, is connected to the bus 60 to indicate to the computer 50 an oxygen burner of the flue gases. A temperature sensor 54 arranged at the entrance of the trap 19 is connected to the bus 60 to indicate to the computer 50 a temperature named T3. input of flue gases in a stage to trap nitrogen oxides. A temperature sensor 5. disposed at the outlet of the stage for trapping the nitrogen oxides, is connected to the bus 60 to indicate to the computer 50 a temperature designated T55 for entering the flue gases in a stage to trap 1-s particles. A temperature sensor 56, disposed at the output of iCtage to trap particles. is connected to the bus 60 to indicate to the calculator 50 a temperature named 1, flue gas at the particle filter outlet. A pressure sensor 59, disposed on or in a cylinder of the engine block 10, is connected to the bus 60 to indicate to the computer 50 a pressure called P ')) engine combustion chamber. It is recalled that a diesel engine usually operates in. lower richness, ie with a fuel oxidant mixture containing fuel in a proportion lower than the stoichiometric proportion or in other words with a superabundance of oxidizing gas. The excess oxygen contained in the oxidant gas which has not been used to oxidize the carbon chains of the fuel during combustion then tends to combine with the nitrogen of the air to form nitrogen oxides. NO and NO2 as a whole denoted by NOx. These NOx are harmful and this is why it is necessary to place them in the flue gases using the NOx trap 19. FIGS. 2a explain an example of a method generally used to purge the NOx trap 19 to maintain its effectiveness. FIG. 2a shows a first graph indicating on the ordinate a richness R of the fuel-fuel mixture in the intake manifold 16 in whip ion of the time t carried on the abscissa. On the left side of the graph, the richness less than 1 corresponds to the normal combustion conditions in the engine. At a time t1, a known software module in the calculator .D gives a richness setpoint R greater than 1, represented in continuous lines in FIG. 2a. FIG. 2b shows a second graph indicating, on the ordinate, a closing rate of the intake flight 61 upstream (the intake manifold 16 again depending on the time t on the abscissa) On the left side of the second graph, The rate of closure of the admission flap 61 equal to Xa °, 1ii corresponds to the normal conditions for obtaining the fuel-fuel mixture At time tl, the known software module in the computer 50 gives, in a phase. 1 a closing instruction of the admission flap 61 equal to Y% higher than X '%, shown in continuous lines in Figure 2b.The closing instruction is established in a manner known per se in the computer 50 to obtain the desired richness R> 1, the admission flap 61 then closes with a response time of its own as shown in dashed line in Figure 2b to reach the closing rate of the same value as the closing instruction at a moment 12 that ends Step 1. The actual richness R then increases during phase 1 as shown in dotted line in Figure 2a. When wealth approaches. of the unit value, the stoichiometric fuel-fuel mixture results in an almost total use of oxygen in the air to crack the hydrocarbon chains of the fuel by a substantially complete oxidation. When the richness R reaches the unit value, the depletion of oxygen causes, during the conversion in the engine, a consequent formation of reducing gases in the flue gases which include, for example, but not just carbon monoxide. The flue gases evacuated in the associated tubing 17 then reach the NOx trap 19 by supplying the reducing gases which react with the nitrogen oxides so as to cause a purging of the NOx trap 19. In a phase 2 which extends beyond the instant t2 until a time t3, the richness R is maintained greater than or equal to the unit value so as to finish reducing the nitrogen oxides trapped in the NOx trap 19. During phase 2, the computer 50 controls an additional injection of fuel generally named post-injection and QP1 amount as shown in Figure 2c. Beyond the moment 13. wealth is reduced to its previous R <1 value to find a traditional mode of combustion. The inventor has observed an appearance of noise when the richness R increases to reach or even exceed the unit value. To study the phenomenon, he noted the evolution of pressure P of a motor cylinder by means of the pressure sensor 59 for crankshaft angles that vary from -1200 to +120 'for the top dead center positioned at D' '. this is a sufficient number of motor cycles to obtain an average which satisfactorily overcomes measurement errors. Sixty-four cycles have yielded good results although a slightly lower number already allows to obtain correct results in terms of lispersion. FIG. 3e shows a change in pressure in the normal mode of operation, in others terms R wealth lower than the unit value. (INC pressure scale plotted on the y-axis on the right of the graph varies from 10 to 120 bar to quantify a first pressure curve in the cylinder.In a first phase of rotation of the crankshaft starting from the bottom dead point to -180 "Up to the top dead center at 0 ', plotted on the x-axis from -120', the piston compresses the oxidant gas in the cylinder A voltage scale plotted on the y-axis on the left of the graph varies from -40 V to 30 \ T to quantify injector controls in the cylinder under consideration, shortly after -60 'a first control of substantially 20 V opens the injector to inject a dose of fuel into 1 cylinder. shortly before -50 ', a -20 V control closes the injector, resulting in a first maximum pressure of about 5 bar just before 0. To -10 °, 25 a second command substantially 20 V opens the injector to inject a new dose of c Flaring in the cylinder .. At 0 ', a -20 V control closes the injector. The resulting combustion causes a second maximum pressure of the order of 105 bars to 10 'of angle after top dead center. In a second phase of rotation of the crankshaft from the top dead center at 00 to the bottom dead point of 180 °, plotted on the abscissa axis up to 120 °, the expansion of the burned males in the cylinder causes a work. recoverable for the propulsion of the vehicle. FIG. 3b shows a pressure evolution in operation mode at richness R greater than or equal to the unit value.

As in FIG. 3a, the pressure scale plotted on the ordinate axis on the right of the graph varies from 10 to 120 bars to quantify a second pressure curve in the liner. In a first phase of crankshaft rotation from low dead point at -180-D to high dead point at 0 ', again reported on the x-axis from -120' the piston compresses the oxidant gas in the cylinder. The voltage scale plotted on the y-axis on the left of the graph varies from -40 V to 30 V to quantify likewise the injector controls in the cylinder considered. Later than in the operating mode shown in FIG. 3a, towards -50 'first control substantially of 20 V opens the injector to inject a dose of fuel into the cylinder. Just before 40, a -20 V control closes the injector. The resultant combustion causes the pressure to rise more slowly than the operating mode shown in Figure 3a, shortly after -20 "the second substantially 20V control opens the injector to inject the new fuel dose. In the cylinder cylinder, shortly before 0 °, the control of -20 V closes the injector, the resulting combustion causes a maximum pressure of about 120 bars to 5 degrees of angle after the top dead joint with a slope. steep as in the figure In the second phase of rotation of the crankshaft from the top dead center at 0c) to the bottom dead point at 180 'plotted on the abscissa axis up to 120 ", the expansion of the flue gases in the cylinder causes here also the work recoverable for the propulsion of the vehicle.

A third control of approximately 10 V just before 60 ° opens the injector to inject a dose of post-injection fuel into the cylinder, and at 60 °, a control of -10 V closes the injector. 3B, the peak pressure is narrower and higher than in Figure 3a, which is a source of shock and therefore of noise in richness greater than or equal to the unit value.The inventor was inulressed auxrets procured on this noise by glazing opening and closing of the valve 62 or the valve 64 recireulation of flue gas, in other words the effects -in: i-és on this noise by different rates of recireulation gases (l Exhaust (EGR) IO The observed evolution of the noise measured for different EGR levels is shown in double line in Figure 1. The EGR rate, that is to say the reeireulation of the gases resulting from the combustion in the cylinders of the engine, is plotted on the abscissa. x EGR rates that vary between a minimum value of Xr% and a maximum value of Yr °, 4i, The purpose of e Figure 4 is to show the appearance of variations of different parameters, which are generally common to many types of engines, without it being necessary here to quantify precisely the values which may be different from one type of engine 20 to another, or from one type of vehicle to another on which the engine is installed, in particular in terms of services a-sound or thermal insulation. The double line curve continues down the graph, shows a pronounced decay of the noise in ric.he.se greater than or equal to the unit value when increasing the rate of EGR by further opening the valve 62 recirculation gas flanges in high pressure. The broken double-line curve at the bottom of the graph shows a less pronounced decrease in higher richness noise or unit value when the rate is increased by further opening the recirculation valve 64 of g "7: burned at low pressure. In view of these curves, there may be a tendency to open up the valve 62 for recirculating the flue gases at high pressure as much as possible, but the opening of the Ffil valves has other effects than simply reducing the noise in richness. greater than or equal to the unit value riotamment that disadvantage the combustion in the engine cylinders by introducing flue gas into a strechiometrique mixture.A drop in temperature in the exhaust pipe then results from the combustion difficulties in the engine. decreasing in single continuous line at the top of the graph of Figure 4, represents the pace of temperature measurements taken by the temperature sensor 5 is placed at the inlet of the trap 19 when increasing the gas recirculation rate at high pressure.

The more strongly decreasing curve in single dotted line also at the top of the graph of FIG. 4 represents the rate of temperature measurements taken by the temperature sensor 54 disposed in the range of the trap 19 when the gas recirculation rate is increased by low pressure.

This decrease in the temperature upstream of the trap 19 is in favor of an increase in the rate of recirculation exhaust gas en it avoids t. erosion that could result from too high a temperature. However, the combustion made more difficult by the increase of the EGR re-circulation rate, also has the effect of increasing the quantity of Cimbrids which are highly reducing such as, for example, the carbon nionoxide CO. The higher the amount of reducing gases entering the trap 19 than the opening of the valve 62 or the valve 64 is large, the effect is to further activate the oxidation-reduction reaction in which intervene the nitrogen oxides that permeate the trap 19, As we know, this oxidation reaction is strongly exothermic. This then results in a temperature increase which is all the more marked as the U 'FOR level is high. The increasing line curve shunts continuously in the center of the graph of station 4, rept. the temperature measurements taken by the temperature sensor 55 disposed inside the trap 19 are increased when the re-circulation rate of 1.sub.az is increased at high pressure. The steeper single dotted curve also in the center of the graph of FIG. 4 represents the rate of temperature measurements taken by the temperature sensor 55 disposed within the trap 19 when the rate of increase is increased. recirculation of gas at low pressure. A good EGI value is therefore the one that achieves the best compromise in ln.) A satisfactory noise reduction and a nondestructive temperature increase, in the trap 19. The good value of the EGR rate is predetermined test phase prejable to mass production of the vehicle or engine. During the test phase, a typical roto-motor is brought to a succession of rotational speeds N varying, for example, from 1000 rpm to a maximum speed of rotation Xi rpm in steps of, for example, 500 rpm. / min. For each rotational speed N, the motor is brought to a succession of effective average pressures, the igE varying, for example, a unit value up to a maximum effective mean ion average Yi. It will be recalled that the effective mean pressure represents a ratio between a work supplied by the engine during a cycle and the displacement of the engine. For each operating point at a given rotational speed and at a given effective average pressure, the EGR level is adjusted to obtain the best compromise in terms of different acceptability criteria for a high temperature in trap 19 and FIG. 'a high level of noise. The value of one or more parameters representative of the EGR rate for the given operating point is then recorded. The test phase is terminated by storing the representative parenthesis value (s) in a table addressed by the rotational speed values N and the effective mean pressure. This double entry table is often called cartouraphic. This table is stored in the computer 50 of each vehicle produced in series so that the computer uses in series the values of the table as servocontrol instructions u the opening of the valve 62 or the valve 64 recirculation burns in function of the rotation speed and the average effective pressure of the engine. The servo set point could for example simply be an opening value of the valve 62 or 64 read during the test phase. However, in addition to requiring a sensor of the degree of opening of the valve, this solution is acceptable but is not entirely satisfactory insofar as the implication c 1 opening on the temperature of the trap 19 and / or on the noise level is impaetée by the environmental conditions of the vehicle such as the outside temperature and / or the speed of the vehicle.

As the desired result is termed - temperature levels and. noise, another solution would be to take a servo set equal to the temperature or noise level detected during the test phase. However, although this solution is also acceptable, a setting on a noise level may induce a temperature drift and vice versa. There are also control loops on several instructions but they are often complicated to implement. During tests, the inventor has noticed that a control value equal to an energy release, ie cumulative, gave a good reproduction in terms of noise temperature regulation.

It is recalled that the energy release Q, in the sense usually accepted in the technical field considered, for an index id angle of the crankshaft, is given by the formula In which the value n in subscript is the angular interval between two successive positions of the crankshaft for each of which the calculator 50 samples, index position j, a pressure given by the sensor 59 and calculates a volume V of gas in the cylinder which is a known function of the geometry of the cylinder and the The formula above is derived from the known laws of thermodynamics, about which it is recalled that. the value of the polytropic coefficient is equal to unity for an isothermal perfect gas transformation, in which case constant PV is obtained. When the. If the gas is of air, the polytropic coefficient generally reaches a value of 7/5 for an adiabatic transformation. For a transformation which is perfectly neither isothermal nor adiabatic, the value of the polytropic coefficient lies between these two values. In a diesel engine, the po ytropic coefficient y generally varies from 1.24 in inferior richness to substantially 1.37 in richness 1. The constant K takes into account the specificities of the engine and the scaling of the engines. units used. It is easily determined by the trade honuno according to the engine under consideration. The calculator 50 calculates a cumulative energy release Q, SM a cycle by summing the instantaneous energy releases Qi over two consecutive engine revolutions. The figure shows that the cumulative energy degeneration on a cycle decreases. almost linearly when the ECIR recirculation tau is used. Thus, the cumulative energy constitutes, in consequence, a state variable of the motor that is suitable for achieving a linear servocontrol of the opening of the valve 62 or of the valve 64. The figures On and Ob show mappings stored in memory of the computer respectively for the high-pressure recirculation valve 62 and the low-pressure re-circulation valve 64. A computer program is installed on the computer 50 to implement the method now explained with reference to FIG. carburation of the engine passes in wealth greater than or equal to 1 reported in step 100, a step 101 consists in determining an engine state variable setpoint from the stored mapping as shown in FIG. 6a to act on the valve 62 recireulation in high pressure or in FIG. 61) to act on the low-pressure recirculating valve 64. As explained above, the state variable is preferably the amount of released energy Qcal calculated from. the pressure P measured by the sensor 50 It will be understood that the method is also applicable to a measured state variable. The setpoint of state variable Qcons is determined at a current rotation speed x = 1 \ 1 and at an average pressure, effective y-PME for example by linear approximation of the Qxy values of the map for values xu, xl and yu , y, cotre which are respectively x and Y. An error F. is evaluated in step 102 by difference between the calculated or measured state variable Qeal and the Qeons state variable setpoint. The error c is used in step 105 to develop an opening command Cr of the valve 62 or 64. The command Cr is for example but not necessarily elaborated by multiplying the error c by a proportional gain A and / or by an integral gain B, the latter resenting the interest of canceling the error in steady state. Note that, in the case where the state variable decode when the opening of the valve 62 or 64 increases as shown in Figure 5 for a state variable equal to the amount of energy released, the Cr command thus developed converts the calculated or measured Qeal constant to the Qcons state variable setpoint. In the error c is positive when () cal is greater than Qcons, thus increasing the opening command Cr in step 105 with the effect of bringing the calculated or measured state variable Que towards the state variable setpoint ( ) cons. Similarly, the error c is negative when Qcal is less than Qcons, thereby decreasing the opening command Cf in step 105 with the effect of bringing the calculated or measured state variable Qcal back to the state variable setpoint Qcons . It is easily understood that for a state variable increasing as a function of the opening of the recirculating valve, it is necessary to reverse the sign of the command Cr in step 105. This would be the case, for example, for the temperature of the NOx trap. 19.

The above explanations remain valid as long as the valves are working correctly. However, the use of the valves 62 or 64 to recirculate the gases when the engine operates in richness greater than or equal to the unit value presents certain risks due to the abundance of imbruls in the gases which can be a source of seizure or clogging of the valves. A malfunction of one or other of the valves 62 or 64 results in an impossibility of making the calculated or measured state variable Qcal coincide with the state variable setpoint Quons within a tolerance interval [cmin , Emax] around ai: lro. The values F.min and cmax can be set equal or different in absolute value during the debugging tests according to the evaluation of the linker (the test.) A step 104 consists in checking that the error e is in the tolerance interval mentioned above As long as these cases, the step 105 is executed in the usual way, a detection in step 101 of the error F. outside the tolerance interval, that is to say less than the negative value Emin or higher than the positive value ElrldX, activates a step t06 which consists of signaling a fault of the DC system, exhaust gas circulation EGR. Step 104 can be continued by activating step 105 to attempt to reduce the error, or by other steps such as inhibiting recirculation of the GR in carburization mode 1, or as going out. 1. In order to avoid an inadvertent activation of step 106 in the event of a transient overrun of the error, a step 103 of filtraee can be provided to produce a filtered error Er with a time constant T.

The filtered error Er is then used in step 104 instead of the error e. It is also possible to use the filtered error Er instead of the error c in step 105, but then taking care to choose the time constant T to avoid in a known manner the instabilities related to the second order transfer functions. .

Step 105 then reboots to step 100 so as to cyclically re-execute steps 101 to 105 as long as your richness is greater than or equal to the unit value.

Claims (10)

REVENDICATIONS1. A method of diagnosing a gas recirculation (EGR) system in an internal combustion engine comprising the steps of: - evaluating (102) an error (by subtracting a value of state value (Qcons) from a calculated or measured state value (Qcal) when the engine is operating in higher richness or equal to a unit value; developing (105) a recirculation valve control (Cr) (62, 64) so as to reduce said fault ; signal (106) a gas recirculation system (EGR) fault if said error (g) comes out of a tolerance interval t Emin, cmax]). 15 The diagnostic method according to claim 1, characterized in that said state value (Qeal) is calculated or measured from a pressure (P) within a motor block cylinder (10). 20 3. Diagnostic method according to one of claims 1 or 2, characterized in CC that said state value (Qcal) is equal to a cumulated energy release (QxyBP, (IP) per motor cycle. 4. A diagnostic method according to I ulie preceding claims, characterized in that said set value of et-t (Qcons) is stored in nu minus a memory cell c.alculatéur (.50) motor vehicle so perform 1,111 on-board diagnosis of the gas recirculation system (liGR), 30 5. Diagnostic method according to one of the preceding claims, characterized in that it comprises a step of determining (101) said setpoint of the state variable (Qcons) of the engine from a map of values (QxyBP , QxyHP) each predetermined for a rotation value (N) and for a mean pressure in 'etive (PME) of the engine in richness greater than or equal to the unit value. 6. The method of diagnosis according to one of the preceding claims, characterized in that it comprises a step consisting of said predetermined state variable setpoint (Qcons) in the test phase of manicie to obtain good acoustic control of the engine. 10 7. Diagnostic method according to one of the preceding claims, characterized in that it comprises a step of predetermined said state variable setpoint (Qcons) in the test phase so as to obtain a good thermal reaction control. chemical reduction of oxides (nitrogen (N0x) produced by the engine. A computer program comprising program code instructions for executing the steps of the method according to one of claims 1 to 7 when said program is executed on a computer. 20 9. Gas recirculation system (Kift) in an internal combustion engine having at least one recirculation valve (62, 64), a motor cylinder cylinder pressure sensor (10) and a calculator (50) comprising a computer program according to claim 8. 10. A gas recirculation system according to claim 9, characterized in that said calculator (50) comprises a state variable setpoint table (Qcons) indexed by a motor rotation speed (N) and / or by an average effective pressure (PI'vrE).