Method for real-time determination of fuel injector flow characteristic

The invention relates to a method for determining in real time, and as a function of the duration of injection control, the performance characteristics of at least one fuel injector, type with electric control, feeding an internal combustion engine and mounted in an engine fuel supply system, this system includes at least one pump, powered from the fuel tank and connected to a common fuel supply manifold for the fuel of engine injectors, each injector is controlled by an engine control assembly comprising at least one computer and at least one memory, generally implemented in the form of an electronic engine control and control assembly, so that in each engine cycle, each injector delivers to said engine a mass of fuel defined by the injector performance characteristics, expressing the injected mass in the increasing function of the duration of injection control of said injector, controlled by the engine control unit, which may also take into account other parameters affecting said characteristics, such as fuel pressure or supply voltage.

[0002] Injectors of this type have a performance characteristic that includes an approximately linear area with injection control duration values ​​greater than the minimum time that is determined by the gain corresponding to its slope and by the initial shift or the shift corresponding to the minimum time duration of control for zero injected mass and obtained at the intersection of the linear extension, until the beginning of the duration of injection control, with an abscissa, expressing the duration of injection control, on a flat graph in which the injected masses are determined on the ordinate axis, as well as the non-linear area of ​​the characteristic, for small values ​​of injection control duration, between offset and linear area. The local reinforcement thus corresponds to the local slope at each point of the curve representing the injector performance characteristics. [0003] Injectors of this type are generally evaluated by their designer on the basis of the theoretical or nominal performance characteristics, whose linear and non-linear theoretical, nominal areas are pre-stored in the engine control unit, for example in the form of theoretical shift and theoretical gain for the linear area, and at least one cartography or theoretical table or mathematical relationship for a non-linear area.

[0004] Fuel injectors with electric control of this type can be fitted with diesel engines or engines with ignition control, and can be installed in direct or indirect injection fuel systems with or without fuel return from the pump before the pump.

[0005] It is known that the injectors used for the injection of the amount of fuel predetermined by the engine control unit have spreads and changes over time of their performance characteristics, which in consequence causes that the injection of a given mass of fuel requires control with different duration of injection depending on the controlled injector and aging of the latter. Indeed, the spread of injector characteristics resulting from the manufacturing tolerances of the physical components of the injectors, i.e. from the dimensional and physical spreads, especially the number and diameter (diameters) of the injection holes, their orientation, elasticity characteristics of their springs, etc. ... and changes over time performance characteristics of the injectors caused in particular by the aging of the physical components of the injectors.

[0006] Furthermore, most of the injection control and control systems, direct or indirect, which are fitted to car combustion engines, ensure that the mixture is controlled in a closed loop and in a continuous manner during engine operation, by means of a probe called the probe λ, which detects the oxygen content in the engine exhaust gas and connected to the engine control unit in such a way as to ensure perfect dosing of the air-fuel mixture, especially when trifunctional catalysts are used for which stoichiometric dosing is required. This closed loop mixture control allows you to satisfactorily compensate for the dispersion of all components that affect the determination of air-fuel dosage and which would affect the parameters for controlling emissions in the engine exhaust gas if the said dispersion would not be compensated. The components concerned are those that allow calculation of the air intake to the engine and control of the fuel injected into the engine so that these elements include the injectors. But, apart from specific strategies, closed loop mix controls do not allow identifying the characteristics of each of the components involved, whether in a holistic or individual way. In other words, the dosing of the air-fuel mixture consists in controlling the efficiency of the air sucked into the engine and the corresponding fuel efficiency and controlling the composition of the mixture in a closed loop enabling compensation of the ratio of air efficiency to fuel efficiency, without identifying the part of the correction relating to air efficiency or fuel efficiency, and furthermore, these mix composition checks do not allow the calculation of an individual correction for each cylinder, so every injector.

[0007] The problem underlying the invention is therefore, starting from the knowledge of the theoretical or nominal performance of the injector, to determine in real time and as a function of the duration of injection control, a change in this characteristic of at least one engine fuel injector to make a shaped relationship occurring between the mass of the injected fuel and the duration of the injection control of at least one considered injector, during shaping phases that occur regularly, during the applicable engine performance characteristics, at action points that are not necessarily in stabilized mode, and during shaping periods short enough not to cause significant deterioration of emissions or perceptible unpleasantness to vehicle users.

[0008] This shaping can relate not only to the performance characteristics of each injector in the same engine, but also to the average or overall set of injectors of the engine under consideration, based on general theoretical or nominal characteristics, determined by general theoretical or nominal reinforcement and general theoretical or nominal shift , as well as across the general theoretical or nominal nonlinear area.

[0009] The object of the invention is, therefore, to allow a better knowledge of the performance characteristics of at least one injector operating an engine by determining in real time the offset of a given injector assuming that its gain is known, and the non-linear part of its characteristics to monitor changes in the individual performance characteristics of each injector, as well as be able to track changes in the overall characteristics of all engine injectors.

[0010] In order to prevent these drawbacks, the method according to the invention, determining in real time and in the function of injection control duration, the performance characteristics of at least one fuel injector with electric control, supplying the internal combustion engine and mounted in the fuel supply system of the type described above, is characterized by that it includes at least stages of admission, that the gain is equal to the theoretical gain or updated gain based on the theoretical gain, and for each injector for which you want to determine the characteristics, the replacement of at least one reference injection, with the duration of injection control controlled by the engine control unit according to the stored characteristic, by injection multiple containing a sequence of at least two injections, whose injection control durations are such that they will result in the injection of the same fuel mass as the swapped reference injection, to determine the fuel mass deviation between the swapped reference injection and multiple injection, to determine from this error the determination of said characteristic and to modify the gain and / or shift of the linear area or at least one array or mathematical relationship of a nonlinear area, to compensate for this error and to remember the new characteristic thus determined.

[0011] According to the invention, it is therefore assumed that the gain of the individual injector under consideration or the general gain of all injectors is considered to be known, since this parameter is subject to small fluctuations and / or because there is already a way to correct this parameter, e.g. known strategies or the method described in French patent application FR 03 02468 filed by

Applicant.

[0012] Replacement of each of the injections with a given duration of injection control by a sequence of two or more injections, the sum of which is to cause the injection of the same amount of fuel, to cause a deviation in the delivered capacity between single and multiple injection, error representative of the injector performance characteristic or injectors, which is remembered in the engine control unit, is a method sometimes sometimes used today, but only for determining the displacement and in the phases of the regulation of the injection system, while the method according to the invention is characterized in that it implements an analogous strategy but in real time, under the conditions allowed or required imposed by the engine control unit, and for determining also other parameter (s) and / or the area (s) of performance characteristics.

[0013] The implementation of the method according to the invention has the advantage that, without the need for manufacturing injectors, and thus their components, with very tight tolerances and thus without increasing the cost of the injection system, ensuring greater accuracy of the mass of fuel injected into each engine cylinder, and as a consequence, ensuring the accuracy of air-fuel dispensing and torque generated by the engine.

This results in good exhaust gas control and greater driving pleasure. In this way, one can be satisfied with equipping the engine with inferior injectors, because the implementation of the method according to the invention makes it possible to compensate for the dispersion at the level of physical elements of the injectors.

[0014] In a preferred embodiment, the method according to the invention comprises at least the steps of first determining the actual offset Or of the linear characteristic area, individual for a single injector or general for the engine injector assembly, replacing each at least one reference injection with duration of injection control T by multiple injection containing a sequence with the number of injections n> 2 with the same duration of injection control

A-ñ⁰- + Ot greater than the minimum time, where Ot is the theoretical or nominal, individual or general shift depending on whether one injector or engine injector assembly is considered and to determine that the actual Or shift is given by the formula Or = Mn-MG + Ot, where Mr and Mr 'are the masses of fuel injected during reference and multiple injections, respectively, n is the number of injections for each multiple injection, G is the gain of the injector or injectors assuming that the real gain is equal to the nominal or theoretical gain (individual or general) remembered in the engine control unit, in which the theoretical or nominal shift Ot is also remembered. [0015] In this way, it is possible to determine the linear region shift of the individual or general characteristic of one injector or all injectors of the same engine, respectively.

[0016] As the fuel mass deviation between swapped and multiple reference injections, i.e. the difference between Mr and Mr ', is proportional to n-1, the fact that the reference injection is swapped by a multiple injection containing a larger number of injections with a shorter control duration makes the shaping method more sensitive to the deviation of the measured performance between two injection modes. However, this increase n quickly meets its limit, because each of the n injections should have a control duration long enough to be greater than the minimum time value, and be in the linear region.

[0017] It is clear, however, that the use of as many n injections as possible is particularly interesting for multiple injections used to determine the shift of the characteristics of one individual injector, because for such a determination, the loss of sensitivity comes from the fact that the deviation of the mass of the injected fuel is then smaller (in relation to m, where m is the number of engine injectors) than the fuel mass deviation specified. when two injection modes (reference injections and multiple injections) are applied to all injectors of the same engine to determine their general characteristics.

[0018] Said fuel mass deviation may, when the engine is of a diesel or controlled ignition type, when the injection system is with direct or indirect injection, and when the fuel supply system is with or without fuel return from behind the pump before the pump, be determined taking attention to the efficiency of the air sucked in by the engine, which is always known to the engine control unit, and the signal from the probe λ to detect oxygen in the exhaust gas of the engine, when the engine injection system is closed-loop controlled and includes such a probe λ located at the outlet of the engine.

As a consequence, in a known manner, the mass of the injected fuel can be measured in each of the two injection modes, reference injection and multiple injection based on the signal generated by the probe λ and the knowledge of the mass of air drawn by the engine, for each operating point of the latter , including if the system is unstable during this procedure.

[0020] In this way, the mass deviation of the injected fuel can be determined taking into account the change in the composition of the air / fuel mixture, based on the signal generated for the engine control unit by the probe λ that detects the oxygen content in the engine exhaust gas and the mass of air consumed by the engine. Especially, the deviation of the mass of the injected fuel can be calculated by the engine control unit based on the probe signal λ and the actual mass of the fuel for injection, said actual mass being determined taking into account the mass of air drawn by the engine and the signal of the actual mixture composition.

[0021] However, when the method according to the invention is carried out in an engine fuel supply system which is a direct injection system in which the common manifold is fed by a high pressure pump which itself is fed by a feed pump connected to the tank, and where the system the supply is a constant volume type and no fuel return from the pump upstream of the high pressure pump, whose performance is controlled and whose model of system behavior is remembered by the engine control unit, it is possible to determine the mass deviation of the injected fuel based on the model of system behavior, according to the change in pressure in the fuel system, caused by forced disruption of pump operation, as presented in the patent FR 2 803 875.

[0022] In this case, it is preferred that the disturbance in the control of the high pressure pump consists in causing the pump to stop, the deviation of the mass of the injected fuel is therefore preferably determined by the pressure drops in the pump supply system that occur after the supply pump stops, on the one hand , when performing at least one reference injection and, on the other hand, when performing replacement injection or multiple injection, where the system behavior model determines the relationship of the mass of injected fuel to each pressure drop.

[0023] After determining the performance characteristics for injection control durations greater than the minimum time, i.e. after performing the shaping of the linear area of ​​this characteristic, the method according to the invention comprises at least the steps of determining the non-linear area and, for this purpose, swapping each at least one reference injection on the duration of injection control T in the said linear area, to which the injected mass M corresponds, by multiple injection containing a sequence of n> 2 injections giving probably the same mass of injected fuel M as the swapped reference injection and having the same injection time control Tn located in the non-linear area, so as to identify the performance characteristics at the appropriate point in its non-linear area, for which it corresponds to the duration of control of each injection, multiple injection

Tn = ^TWell^r + Or, the mass of the injected fuel M / n, and change ni / or Tn, to identify at least one part of the non-linear area.

[0024] In this case, it is also possible to make the non-linear area of ​​the individual, one considered injector, or general, if all injectors are considered, in which the gain, shift and non-linear area are considered theoretical or nominal individual, or gain, shift, and theoretical or nominal general nonlinear area.

[0025] In a variation, to determine the non-linear area after determining the linear area, the method according to the invention may consist of imposing the duration of injection control T2 in the non-linear area for which the characteristic is to be determined, and replacing each at least one reference injection with duration injection control T in the linear area, to which the injected mass M corresponds, by multiple injection on> 2 injections, which is to give the same mass M of injected fuel as the swapped reference injection and where one has the duration of injection control T1 located in the linear area and to which the injected mass M1 corresponds and where the remaining n-1 injections have each the same imposed duration of injection control

T2, and each of them corresponds to the injected mass M2 such that M2 = i on the differentiation of T2, smaller than the minimum time, and / or n so as to determine at least part of the non-linear area.

[0026] In this variation also, the non-linear area can be defined generally for all injectors of the same engine, either individually, for one or each of them, but in this case, with less sensitivity.

[0027] Also, in two modes of carrying out the method of the invention for determining the non-linear area, determining the mass deviation of the injected fuel Mr-Mr 'or M-M1, as mentioned above, can be obtained either based on the efficiency of the air sucked in by the engine and the signal from probe λ, or in the special case of the system without fuel return from behind the pump to the pump, in the direct injection system in which the engine control unit knows the model of the system behavior, and controls pump performance, controlling pump stop and measuring pressure drops during reference injections and multiple injections, respectively, to determine the mass of fuel injected, via a system behavior model.

[0028] As mentioned above, the method according to the invention can be applied simultaneously to either all engine injectors or to one injector, in which case the method is preferably used sequentially for each engine injector in such a way as to shape their individual characteristics.

[0029] Other features and advantages of the invention will become apparent when reading the following description given by way of non-limiting embodiment shown with reference to the accompanying drawings, in which:

- figure 1 is a diagram of the fuel supply system of a car internal combustion engine by direct injection to implement the method according to the invention,

- figure 2 shows the performance characteristic, which can be the general characteristic of all the injectors of the system in figure 1, or the individual characteristic of a single injector,

- figure 3 shows the change in pressure in the common manifold of the system of figure 1, as a function of time, in the case of two pressure drops caused by stopping the pump of the system of figure 1, each of which is obtained respectively for one of two different injection modes, controlled for a certain number injections in all injectors or in one of them,

- Figure 4 is a set of three characteristics that simplify the replacement of the reference injection (Figure 4a) in the linear region, by multiple injection with two successive injections, one of which is in the linear region (Figure 4b) and the other in the non-linear region (Fig. 4c),

- figure 5 shows a set of four characteristics, one of which for reference injection in the linear region (fig. 5a) and the other three for multiple injection with three consecutive injections, one of which is in the linear region (fig. 5b) and the other two in at the same point of the non-linear region (Fig. 5c and Fig. 5d), and

- figure 6 is an analogous figure to figures 4 and 5, showing a set of four characteristics, one of which corresponds to the reference injection in the linear region (Fig. 6a) and the other three for multiple injection with three successive injections at the same point of the non-linear region (Fig. 6b, 6c and 6d).

[0030] In Figure 1 the combustion engine 1 for a car is schematically represented. For example, engine 1 under consideration is a four-cylinder in-line engine with a controlled ignition and four-stroke engine cycle fueled by direct injection, although the method of the invention can be used in an indirect injection engine and / or a diesel engine type. [0031] Fuel injection is provided in each engine cylinder 1 by one of four injectors respectively

2.

[0032] These injectors 2 are supplied with high pressure fuel through a common fuel manifold 3 in which the fuel pressure is determined, at least at certain points in the engine cycle, by measuring by a pressure sensor 4 sending a measured pressure signal to the engine control unit 5, or by calculating in this assembly 5 based on certain measurements made by sensor 4 at certain times of the engine cycle, as proposed in patent FR 2 803 875.

[0033] The engine control assembly 5 is the electronic control unit for fuel injection in engine 1, controlling through the wiring harness control 6 the moments and times of controlling the injection of the injectors 2, as well as the ignition in the cylinders of the engine 1, in the example of a controlled ignition engine under consideration, and possibly other functions such as control of air intake to the engine via a motor-driven throttle, in particular, the accelerator pedal function and other safety functions such as anti-slipping and / or anti-lock braking. This electronic assembly 5 comprises, in a known manner, at least a computer, in particular with computing, memory and comparison systems, and in the injection control function, the assembly 5 for controlling and controlling the amount of fuel injected by each injector 2 into the respective engine cylinder 1, as a function of engine strokes in each cylinder, engine parameters and operating conditions, especially in its load system, or also temperature and fuel demand, in function of especially the efficiency of the air sucked into the engine 1 and the moment that the engine should develop, these parameters are entered via connector 7 to the engine control unit.

[0034] In this example, the common manifold 3 is supplied with high pressure fuel through a high pressure pump 8, with controlled capacity and connected to the manifold 3 through a pipe 9 in which the fuel flows in the direction of arrow F1 and the engine control unit 5 controls the high pressure pump 8 through a logical connection 10 and the mass of fuel thus determined is fed by the high pressure pump 8 to the collector 3, in each cycle of engine 1.

[0035] The high pressure pump 8 is rotatably driven in a known manner by the engine 1 via a mechanical connection shown as 11. The high pressure pump 8 is itself supplied with fuel through a delivery system comprising, from start to finish, a fuel tank 12, a feed pump or low pressure pump 13, submerged in tank 12 and fed through a filter (not shown) and a fuel pressure regulator 14, the output of which allows the return of excess fuel to the tank 12, and whose second output is connected to the suction inlet of the high pressure pump 8, at level, on which is located an electrically controlled valve (not shown) controlled in the system all or nothing from the assembly 5 through a logical connection 10, so that the fuel capacity of the high pressure pump 8 is known to the control assembly 5 that can control this electrically controlled inlet valve in such way to force the high pressure pump 8 to run at zero.

[0036] The system of supplying the engine 1 with fuel by direct injection is thus a high pressure system, comprising a high pressure pump 8 and components below the latter, i.e. a pipe 9 and a common manifold 3, this high pressure system which is a system of constant volume and without constant fuel return or without fuel recirculation from behind the pump upstream of the high pressure pump 8, it is fed by a low pressure feeding system, upstream of the high pressure pump 8, containing tank 12, pump 13 and regulator 14.

[0037] In this way, the mass of fuel in the high pressure system results only from the filling operation by the high pressure pump 8 and the fuel injection into the engine 1 by the injectors 2, and these activities are controlled by the assembly 5.

[0038] Performance characteristics of one injector

2, expressing the mass of fuel injected Mi as a function of the duration of the Tinj injection control, as determined by the assembly

5, corresponds to the increasing function, the graph of which is shown in Figure 2, and with a slope equal to the local gain G of the injector, which is related to the entire value of the injection duration and determined by the ratio of the change in injected mass caused by a small change in the duration of the injection, and the change in time itself injection duration. This graph contains an approximately linear area 15 in which the G gain is constant and a nonlinear area 16 with low values ​​of injection control duration (values ​​smaller than the minimum time corresponding to the lower linearity TinfL) and in which the local gain changes quickly.

[0039] The linear region 15 of the characteristic is determined not only by its slope or the constant gain G of the injector in this area, but also by the initial shift or shift Ot, at the intersection of the extension of the linear part of the curve to the beginning, with the abscissa indicating the duration of the control Tinj injection.

[0040] It is known that the MinfL mass, which is injected during the duration of injection control equal to the lower limit TinfL of the linear area 15 is equal to the sum of the injected masses during the transient phases corresponding to the phases of determining and cutting off the instantaneous efficiency of one injector 2 caused respectively by the opening and closing of the injector 2 resulting from the displacement of the injector valve accordingly setting and cutting off the excitation current of the injector coil with electromagnetic control, which occurs at the beginning and end respectively of the logical injection control command developed in assembly 5 and sent by the latter to the given injector 2 through the appropriate wire from the bundle 6.

[0041] Generally, injectors 2 of the same type are evaluated on the basis of the theoretical characteristics of the injector efficiency, determined, on the one hand, by the theoretical amplification Gt and the theoretical offset Ot, to determine the theoretical linear area of ​​the graph, and on the other hand, by the theoretical non-linear area 16 resulting from the application of one or several mathematical dependencies and / or stored in assembly 5 in the form of tables or cartography indicating the injection mass Mi for the duration of injection control Tinj contained between the lower limit of linearity TinfL and the theoretical shift Ot and in the range of injection control time corresponding to the non-linear area 16 .

[0042] Starting from this theoretical characteristic, which is individual (for one injector 2) or general (for all injectors 2), the method according to the invention aims to determine in real time (engine 1 running) this characteristic or individual (for one and preferably each of the subsequent injectors 2), or general (for all injectors 2 of engine 1), starting with shaping the linear area of ​​the characteristic, and for this purpose, it is assumed that the gain G is constant and remains equal to the theoretical gain Gt, or the gain is updated on the basis of theoretical gain, for example by carrying out the method described in French patent application FR 03 02468 filed by the Applicant. G gain can be considered constant because its value is not very susceptible to fluctuations.

[0043] Consequently, determining the linear region of the characteristic provides for determining the Or shift of this linear region.

[0044] To this end, during the shaping time interval, the engine control assembly 5 controls the swapping, for example in all injectors 2, if one wants to determine the actual overall offset of the linear area of ​​the overall performance characteristics of the injectors 2, a number of injections called reference, having times duration of injection control located in the linear area of ​​theoretical and nominal characteristics, and corresponding to the engine 1 requirements for those engine operating points, such as those defined by assembly 5, by the same number of multiple injections, each of which forms a sequence of at least two injections, whose duration of control of each injection is greater than the minimum time, i.e. also located in the linear region of the output characteristic 15. Typically, each reference injection is replaced by a multiple injection consisting of two successive injections for which the actual duration of each injection, i.e. the duration of its injection control reduced by a known offset, i.e. theoretical offset

Ot, it is equal to half the actual duration of the injection, reference injection, so that it is assumed that two successive multiple injection injections inject the same mass of fuel into the engine as the swapped reference injection.

[0045] In other words, in this case, each reference injection, with an injection duration of T equal to (in the linear region, and thus greater than the minimum time) is replaced by a multiple injection consisting of two successive injections having each duration of injection control equal to ^T—^Ot + Ot, while the mass of Mr fuel actually injected by reference injection or the injected reference mass is given by the formula Mr = G x (T - Or), where Or is the actual sought shift, while the mass of Mr 'injected by two successive injections multiple injection can be expressed by the following formula: Mr '= 2.G. (^TfO + Ot - Or) in which Mr and Mr 'are injected masses during reference injection and during multiple injection (in this case double), respectively,

G is the general gain of injector 2 (assuming that this gain is equal to the general theoretical gain),

Or is the actual overall shift of the general characteristics of injector 2,

Ot is the theoretical general shift remembered in group 5, and

T is, as already mentioned, the duration of the injection control, the reference injection selected in the linear region.

[0046] From the above formula expressing Mr i, it is obtained that the actual offset G gain is combined by the following

Or = Ot + ^mrG^MLG

Mr ', Or and formula:

[0047] More generally, if each reference injection (with the duration of injection control T in the linear region) given to injectors 2 during the shaping time is replaced by multiple injection on> 2 injections, where the duration of control of each injection is sufficiently large, in order to be greater than the minimum time Tinf L, each of the n multiple injection molding therefore has the duration of injection control equal + Ot. The mass of fuel injected with each multiple injection is therefore equal to: Mr '= η (Ά-^τ- ° - + Ot - Or). [0048] The actual shift is therefore calculated from the following formula: Or =^Μ~^Μ'+ Ot [0049] Since the difference between Mr and Mr' is proportional to (n-1), the fact of swapping each reference injection by multiple injection created by a larger number of shorter injections makes the shaping method more sensitive to the deviation of the mass of injected fuel measured between the two modes injection (reference injection and multiple injection). [0050] The method of determining the actual offset, as described above, can only be used for one injector 2, taking into account the theoretical, individual offset and individual gain of this injector 2, so as to determine the actual individual offset. But it is understandable that this determination is carried out with a certain loss of sensitivity, because the deviation between the masses of injected fuel, on the one hand by making reference injections and on the other hand by performing multiple injections replaced in the considered injector, will be smaller than for determining the actual overall shift and in the first roughly you can estimate that the deviation of the injected mass between two injection modes for one injector 2 is equal to the ratio of this deviation calculated to determine the actual overall shift at the number of m injectors, doing the shaping with the same number of reference injections and swapped and the same swapped multiple injections.

[0051] So, to know the actual overall or individual shift, the mass deviation of the injected Mr-Mr 'fuel must be determined.

[0052] This determination of the mass deviation of the injected fuel can be made according to changes in the composition of the air-fuel mixture and knowledge of the mass of air drawn in by the engine 1, in virtually all injection systems, with direct or indirect injection, in gasoline or diesel engines and with fuel supply systems , which are not necessarily of the special type described above, that is to say, of a constant volume, with a pump with controlled capacity and no fuel return from behind the pump in front of the pump and in which the model of the system behavior is known to the engine control unit 5, and with the proviso that these injection systems have a closed loop mixture control, provided by a probe λ 17, located in line the exhaust 18 of the engine 1, and detecting the oxygen content in the exhaust gas, this probe λ is connected to the assembly 5 to send signals to it. [0053] In a known manner, the determination of the mass of injected fuel during each of the two mentioned injection modes may consist in the division of the mass of air taken by the engine during each of these phases and measured by the control unit 5 by a factor proportional to the factor λ, which is itself measured by probe λ 17 or calculated according to the signal from this probe, according to the formula: λ =) AFg in which A and F are air and fuel measurements, respectively, and the index s corresponds to the stoichiometric value of the A / F ratio.

It is known that the coefficient λ can be measured directly if the engine 1 is equipped with a proportional probe λ 17 in the exhaust system 18, and that this coefficient λ can be determined from the correction value that the closed loop composition of the mixture performs if the system contains a probe type λ 17 all or nothing (attached-disabled).

[0054] In a manner known to the skilled person, the mass of the injected fuel can be calculated by the assembly 5 on the basis of, on the one hand, the probe signal λ, and on the other hand, the amount or actual mass of the injection fuel, which is itself determined taking into account the actual composition signal mix and actual air mass calculated by assembly 5.

[0055] By way of example, with a stabilized nominal operation, i.e. after the closed-loop injection control by assembly 5 and probe λ 17 determined the value of the mass of fuel injected to the actual mass determined in assembly 5, and the automatic correction determined the mean value of the correction factor λ closed loop, the whole deviation of the coefficient λ following the execution of special injection control, that is, performing multiple injections instead of reference injections, it represents the change in mass of injected fuel equal to Mr '-Mr = M. ^λ-^-λ , wherein:

Mr 'is the mass of fuel injected during multiple injections,

Mr is the mass of injected fuel, at the same point of engine operation, when making reference injections (without using special control),

M is the value of the actual mass of fuel, which was calculated by the engine control unit 5 for a given engine operating point, λ is the expected value of the coefficient λ before performing multiple injections (this value is optionally measured if stability conditions allow it), and λ ' is the value of the coefficient λ measured after administration of multiple injections.

[0056] This determination of the weight of the injected fuel mass can also be made if the "injection-fuel-engine" system is not stable during the determination procedure.

[0057] But because the system of figure 1 is a special system, with a constant volume and no fuel return from behind the pump before the pump 8, the capacity of which is controlled, in which the engine control unit 5 has in mind the model behavior of the system, the deviation between the mass of fuel injected during two modes of reference injection feeding, and swapped multiple injection can be measured in a different way, according to changes in pressure in the fuel system, following the disruption introduced to the operation of the feed pump 8, and especially after the feed pump 8 has stopped, on the one hand, during reference injections, and on the other, during multiple injections, based on the system behavior model that corresponds to each drop in the measured pressure of the injected fuel mass , according to the recommendations given in patent FR 2 803 875.

[0058] According to this patent, the relationship between the pressure drop in the manifold 3 and the mass of fuel injected into the engine 1 is provided in assembly 5 by the high pressure supply system maintenance module 18, this module containing the memory in which it is remembered, in the form of tables or cartography , the law specifying the change in fuel mass in a high pressure system as a function of a specific pressure drop in this system when the pump stops 8.

[0059] This measurement of the deviation between the mass of fuel injected during the two mentioned injection modes (reference injections and swapped multiple injections) can be made as described with reference to figure 3, which shows changes in pressure P as a function of time t in a common manifold 3. [0060] Coming out of the state in which the engine 1 is running while the pressure Po prevails in the manifold 3, at the moment the assembly 5 controls the stopping of the pump 8, while the reference injections, with the duration of the injection control in the linear area of ​​the performance characteristics are fed to the injectors 2. Pressure P drops from Po, when pump 8 stops, to pressure P1 at t1, corresponding to the end of the blocking period of pump performance 8, and after a sufficient number of reference injections fed to injectors 2 so that the pressure drop of DP1 = Po-P1 can be measured with sufficient accuracy by sensor 4, this pressure drop DP1 results from the supply of engine 1 cylinders via injectors 2 from collector 3, while this collector 3 is no longer supplied by pump 8.

[0061] Thanks to the high pressure system behavior model memorized in module 18 of assembly 5 and based, for example, on the mass of fuel entering the collector 3 and imposed by the high pressure pump 8, which is determined by the computer 17 of assembly 5 and on the mass flowing out of the collector 3, which is injected into engine 1 and also determined by assembly 5, as well as on the stiffness of the high pressure system, this corresponds to the pressure difference DP1 thus determined, the first mass of fuel injected into the engine 1 by all the injectors 2, and which corresponds to the said mass [0062] After the failure of the operation of the high pressure pump 8, and the return of the normal operation of the engine 1 at a given operating point, a second phase of mass measurement is initiated, which consists in reintroducing the same interference as previously in the operation of the high pressure pump 8, that is, at the interruption of its operation during the interval t1-t0 during which the same number of swapped multiple injections is given as the number of reference injections given during the same period of time t1-t0 leading to a drop in pressure DP1. Performing these swapped multiple injections while the pump capacity is zero, leads from the initial pressure P0, through the pressure drop DP2 to the pressure P2 at time t1. Thanks to module 18 of assembly 5, in which the high pressure supply system behavior model is saved and remembered, the pressure drop DP2 corresponds to the second mass of fuel that flowed out of the high pressure system and was injected by injectors 2 into engine 1, this second mass of fuel is the said mass Mr '.

[0063] The assembly 5 can thus calculate the mass difference of the injected Mr-Mr 'fuel, which allows the actual offset to be calculated. Or.

[0064] The linear area of ​​the overall performance characteristics can thus be updated and stored in the assembly

5.

[0065] To update and remember the linear area of ​​the individual performance characteristics of one injector 2, it is enough to reproduce the process described above using the same normal and / or reference injections during the two phases of injecting to all injectors 2 except the one whose characteristic we want to determine, this injector 2 is the only one in which reference injections were used during the first phase, followed by multiple injections instead during the second phase. Of course, in this case, to obtain the same sensitivity as before, the number of reference injections used and the multiple injections replacing them will be greater, to take into account the fact that the deviation of the mass of the injected fuel is caused only by the participation of a single injector 2.

[0066] It should be noted that the two phases can be inverted, the injected mass Mr 'as a result of the replacement multiple injections is determined before the injected mass Mr, as a result of reference or normal injections, or also a series of non-adjacent two phases can be repeated a number times changing the order of the phases. But, for a good determination of the actual Or offset and the area of ​​the linear performance characteristic, general or individual, this shaping procedure should be repeated for different engine operating points, for a sufficient number of injection control durations in the reference injection linear area and possibly for different injection numbers spare multiple injections.

[0067] The linear performance characteristics, general or individual, have been updated and remembered, based on knowledge of the actual Or shift and G gain, the determination of the non-linear area of ​​this characteristic remains to be done. [0068] It is assumed that a non-linear theoretical or nominal area of ​​theoretical or nominal performance characteristics is known that is remembered in assembly 5. To make the determination of a non-linear, general or individual area, the method according to the invention proposes, after determining the linear area of ​​the performance characteristics, i.e. for the duration of injection control greater than the minimum time, to determine the conditions of use of all injectors 2 (determining the general characteristics) or a single injector 2 (determination of individual characteristics) in a linear area, then dividing each reference injection, in the updated remembered linear area, into a number n, at least equal to two, of the injections of the replacement multiple injection, so that the sum of these n injections is able to give the same mass of injected fuel as the reference injection.

[0069] Three embodiments are described below with reference to figures 4, 5 and 6 respectively, two examples of figures 4 and 5 correspond to the first embodiment and the example of figure 6 to the second embodiment.

[0070] In the first embodiment (figures 4 and 5), if one wants to impose in advance the value of the injection control duration T2 selected in the non-linear area, and for which one wants to know the point corresponding to the real non-linear area or the updated performance characteristic, each one is replaced from a number of reference injections having duration of injection control in the real or updated linear area, for replacement multiple injection, formed by the sequence of n injections, in which the number of n-1 injections has the duration of injection control identical and equal to T2, and in which the nth injection has the duration of injection control T1 (or T'1) in the linear area of ​​the actual or updated characteristic, that the sum (n-1) of injections with increased duration of T2 injection control with the duration of T1 (or T'1) control will result in the injection of the same mass of fuel as a single reference injection with the duration of control T.

[0071] If, for example in the case of general characteristics, the injector assembly 2 has a nominal performance characteristic, the fact of dividing each injection control duration into n injections as described above should not affect the total weight of the injected fuel. But if the mass injected by making multiple replacement injections is different from the expected one, i.e. obtained when making reference injections, the deviation from this expected value is representative of the error of the injector unit 2 performance for the point corresponding to the duration of the T2 injection control selected in the non-linear area, relative to the predicted value, in relation to the theoretical or nominal nonlinear area. By repeating operations for several predefined values ​​of the T2 injection control duration in the non-linear area, it is possible to reconstruct the non-linear area of ​​the general characteristics of the injectors over the entire range of injection control times that are smaller than the minimum time.

[0072] In Figure 4, graph 4a shows the linear area 15 'whose shaping has been made, the performance characteristics whose non-linear area 16 is theoretical or nominal, and the reference injection with the duration of control T in the updated linear area 15' provides mass injection M . This reference injection is replaced by a multiple injection formed by a series of two injections, one of which, shown in diagram 4c, has the duration of injection control T2 selected in the non-linear theoretical or nominal area 16 and which corresponds to the mass of the injected fuel M2, which is to be determined for accurate knowledge of the appropriate point in the updated nonlinear area. The second injection (see graph 4b) of the replacement multiple injection corresponds to the duration of injection control T1 in the updated linear area 15 'and which corresponds to the mass of the injected fuel M1 precisely determined by the shaping made for this linear part 15' of the characteristic. The duration of the T1 control is selected so that the sum of two injections with the control duration T1 and T2 results in the injection of the same mass of fuel M as the reference injection with the duration of the control T from curve 4a.

So M = M1 + M2 is assumed. Where M2 = M - M1. The values ​​of M and M1 are known precisely because determined from the updated linear area 15 'of the characteristic, the exact mass value M2 is obtained, so that the point (T2, M2) of the real nonlinear area is precisely determined.

[0073] The example shown in figure 5 differs from the example described above in relation to figure 4 only by the fact that the reference injection with the duration of control T in the updated linear region corresponding to the mass of the injected fuel M, shown in diagram 5a, is replaced by multiple injection formed by a series of three injections, two of which are presented in graphs 5c and 5d, have each T2 control duration selected in the non-linear theoretical or nominal area and to which the injected mass M2 corresponds, while the third injection is shown in diagram 5b and corresponds to the duration of the control T'1 in the updated linear area 15 'of the characteristic and to which the injected corresponds mass M'1. Since it is assumed that the sum of three injections from Figures 5b, 5c and 5d gives an injection with the same mass of fuel M as a single reference injection with the duration of control T from Figure 5a, so M = M'1 + 2M2 is obtained, from where M2 = [0074] In a more general case, if the replacement multiple injection is formed by a series of n injections, of which (n-1) with the duration of the control T2, and the last with the duration of the control in the updated linear region and to which the injected mass M1 corresponds, the injected mass for each of the (n-1) injections is:

M 2 = ^M-M¹- .

n-1 [0075] In the second embodiment, which for example is described with reference to figure 6, n injections of each replacement multiple injection have the same injection control time Tn selected in the non-linear theoretical or nominal area 16 so that their sum causes injection the same mass of fuel as in the swapped single reference injection, having the duration of injection control T in the updated linear area 15 'of the characteristic and to which the mass of the injected fuel M corresponds exactly known. You can thus identify the updated nonlinear area of ​​the performance curve at the appropriate point, because it corresponds to the duration of the injection control Tn, the mass of the injected fuel m is equal to M / n.

[0076] The example of figure 6 is an example in which each reference injection with control duration T in the updated linear region 15 'and the corresponding injected mass M are replaced by a multiple injection formed by a sequence of three identical injections with the same control duration T3 in the non-linear theoretical or nominal area 16 and to which corresponds to the mass of fuel injected m, so that the sum of these three successive injections of the replacement multiple injection gives an injection of mass equal to M. So M = 3m, from where m = M / 3. This makes it possible to accurately determine this point (T3, m) of the updated non-linear area that can be reproduced by changes of T3, and possibly n (equal to 3 in the example of figure 6).

[0077] It is understood that in this case, the duration of injection control Tn of each injection of the replacement multiple injection is equal to ^T- + Or, because n (Tn-Or) = T-Or, if it is assumed that n injections of replacement multiple injection gives the same mass of fuel as a single reference injection, provided that the G gain is constant and the same for different injections.

[0078] To confirm the values ​​of the injected masses read in the linear areas of the characteristics, or for comparison with the read values, or also instead of these readings, the injected masses or differences of the injected masses can be determined by applying the method described above by introducing the mixture composition factor λ and the mass of the intake air by the engine, or if the structure of the high pressure system makes it possible, by a method introducing a system behavior module that determines the relationship of injected fuel masses and pressure drops measured in the fuel manifold 3, while pump 8 performance is temporarily blocked during two reference injection modes and replacement multiple injections.

[0079] The updated nonlinear area can thus be determined and complement the shaping of the performance characteristics in real time, general or individual, because the actual Or shift has been previously determined and it is assumed that the G gain is constant and equal to the theoretical gain, or it can be updated by another strategy adapted for this purpose.

MAGNETI MARELLI MOTOPROPULSION

FRANCE SAS

Proxy:

EP 1644 627 B1