EP2042222A1 - Device and method for detecting the clogging of a fuel filter of an internal combustion engine fuel supply system - Google Patents

Abstract

The device has a determination module (25a) determining clogging of a fuel filter (16) from an integral component of a proportional integral derivative (PID) regulator type regulation unit (25), when average values of the integral component are higher than respective maximum thresholds. The determination module evaluates the integral component of the adjusting module in a cyclic manner for stabilized functioning phases. An alert module (30) alerts a driver for the detection of the clogging of the fuel filter by the determination module. An independent claim is also included for a method for detecting clogging of a fuel filter of a fuel supply system in an internal combustion engine.

Description

The invention relates to a device and a method for detecting the fouling of a fuel filter of a fuel supply system of an internal combustion engine, particularly a motor vehicle. More specifically, the invention relates to high pressure direct or indirect injection systems, for example of diesel type.

Conventionally, if there is a difference between the desired pressure and the pressure actually measured in the common injection rail of the engine, the actuators, also called control valves, are controlled so that the measured pressure tends to the desired pressure.

The fuel supply systems are designed to achieve increasingly higher injection pressures, for example greater than 1600 bar. Such fuel supply systems require the use of a high purity fuel. Also, it is known to filter the fuel by arranging a fuel filter on the low pressure circuit, between the fuel tank and the pumping assembly.

Over time, the operation of the filter causes fouling or clogging thereof. Clogging of the fuel filter can disrupt the proper functioning of the engine fuel supply system, in particular by causing a pressure drop at the inlet of the pump assembly. Clogging the fuel filter can even interrupt the fuel supply to the engine and cause it to stall.

Also, automakers advocate changing the fuel filter when it participated in a mileage threshold of the vehicle. This solution is expensive because the fuel filter will often be changed too early.

There are systems as described in the International Patent Application WO 2005/098227 (UFI FILTERS) that use pressure sensors to directly measure the pressure drop across the fuel filter, and directly deduce the degree of fouling of the fuel filter. However, the use of pressure sensors at the terminals of the fuel filter has a high cost.

The French patent application FR 2 787 143 (Magneti Marelli France) discloses a method and a system for detecting the fouling of a fuel filter arranged, in a fuel supply circuit of an internal combustion engine, between, on the one hand, a a fuel pressure regulator, downstream of the filter, and of the bypass type, delivering fuel at a pressure imposed upstream and to the internal combustion engine, and on the other hand, a fuel delivery pump from a reservoir, the pump being driven by an electric motor and disposed upstream of the filter for supplying the regulator through the filter. This document discloses a pressure regulation from the difference between the inlet and outlet pressures of the fuel filter. This pressure difference can be measured by sensors, or estimated.

Since the solution consists in using pressure sensors at the input and at the output of the filter being expensive, the document presents a method which makes it possible, on the one hand, to estimate the pressure at the input of the filter, or at the pump output, from less than the instantaneous rotation speed of the pump and the average supply current of the electric drive motor of the pump, and secondly, to assimilate the fuel pressure at the outlet of the filter as the pressure imposed by the pressure regulator.

Such a system necessarily requires an electric motor driving the pump, and a means for estimating the supply current of said electric motor, which is expensive.

The estimation of the pressure at the outlet of the pump from the instantaneous rotation speed of the pump and the average power supply current of the electric drive motor of the pump is done in an open loop through an operating model of the pump, which generates a lack of precision in particular because of manufacturing dispersions and drifts over time of the pump assembly and the electric motor.

The thermal state of the pump must necessarily be taken into account in the operating model of the pump to obtain an improved estimation in open loop of the pressure at the outlet of the pump. However, such an estimate is expensive because it requires either the use of a sensor to measure the temperature of the pump, or the use of a thermal model of the pump, knowing that a thermal model pump is not necessarily easy to calibrate and may also be imprecise.

As for the determination of the inlet pressure of the filter (or pump output), the determination of the pressure at the output of the filter assimilated to the pressure imposed by the pressure regulator is in open loop.

Indeed, regardless of the fuel demand of the engine, the estimate of the pressure at the filter outlet is based on the characteristic curves (including flow-pressure) known to the regulator and stored in the electronic control unit of the engine. Also, this can generate a lack of precision on the actual estimate of the pressure at the outlet of the filter, since the manufacturing dispersions and / or the drifts of the regulator are not corrected.

Finally, a pressure regulation based on a pressure regulator operating as a bypass downstream of the pump generally has a lower energy efficiency than a pressure regulation based on a fuel flow regulator placed upstream of the high-pressure pump. pressure. Indeed, with a bypass operation where the excess fuel is returned to the tank by the pressure regulator, more fuel is compressed than necessary.

The present invention aims to provide a solution to these problems.

An object of the invention is therefore to propose a device for detecting the fouling of a fuel filter of a fuel supply system of an internal combustion engine making it possible, at reduced cost, to determine, with improved accuracy, fouling or clogging of the fuel filter necessitating the change of the fuel filter.

For this purpose, a first aspect of the invention relates to a device for detecting the fouling of a fuel filter of a fuel supply system of an internal combustion engine, especially a motor vehicle. The fuel supply system comprises a pumping assembly, comprising a first low pressure pump and a second high pressure pump arranged in series, and being arranged between a fuel injection ramp provided with a pressure sensor and a fuel tank. The fuel supply system further comprises a first controlled fuel flow control valve supplying the high pressure pump, means for determining a set pressure in the injection manifold, and means for regulating said first controlled valve. The regulating means of said first controlled valve comprise Proportional Integral and Derivative components and a feedback loop for controlling the pressure measured by the sensor on said setpoint pressure. The fuel filter is disposed between the fuel tank and the pumping assembly. The device comprises means for determining the fouling of the fuel filter from said integral component of said regulating means.

By avoiding the use of pressure sensors at the terminals of the fuel filter, it is thus possible to detect, at reduced cost, a fouling of the fuel filter.

In one embodiment, said determination means are adapted to evaluate the integral component of the regulation means for stabilized operating phases.

In one embodiment, said determining means are adapted to carry out said evaluation cyclically, a cycle corresponding, for example, to the consumption of a threshold fuel quantity or the threshold distance of the vehicle. .

It is thus taken into account the rapidity of the phenomenon of fouling of the fuel filter with respect to the wear or aging of any other component of the fuel supply system.

In one embodiment, the determination means are adapted to detect a clogging of the fuel filter when a set of average values of said integral component, depending on the speed of rotation and the flow rate delivered by the first pump, to the cycle of current observation k, are greater than respective maximum thresholds dependent on the speed of rotation and the flow rate delivered by the first pump, the current observation cycle k being greater than or equal to a number of reference cycles n_ref increased by at least two.

In one embodiment, the determining means are adapted to use arithmetic or sliding averages.

dans laquelle
x̅_n est la moyenne des n dernières valeurs,
x̅_n-1 est la moyenne des n-1 dernières valeurs, et
x_n est la dernière valeur calculée.Using a sliding average makes it possible to minimize the amount of data to be memorized. A sliding average is calculated from the average of the previous values, and the last value calculated by the following formula: $${\stackrel{\mathrm{~}}{\mathrm{x}}}_{\mathrm{not}}\mathrm{=}\frac{\left(\mathrm{not}\mathrm{-}\mathrm{1}\right){\stackrel{\mathrm{~}}{\mathrm{x}}}_{\mathrm{not}\mathrm{-}\mathrm{1}}\mathrm{+}{\mathrm{x}}_{\mathrm{not}}}{\mathrm{not}}$$

in which
x̅ _n is the average of the n last values,
x̅ _n-1 is the average of the last n-1 values, and
x _n is the last calculated value.

In one embodiment, the determining means are adapted to detect fouling of the fuel filter when, in addition, the differences between said average values of said integral component, at the current observation cycle k, and corresponding values, for the reference cycle n_ref, depending on the speed of rotation and the flow rate delivered by the first pump, are greater than the corresponding differences for the previous cycle k-1 or the ante-penultimate cycle k-2.

In one embodiment, the determination means are adapted to detect clogging of the fuel filter when, moreover, differences between first differences between mean values of said integral component for a maximum flow delivered by the first pump, to the current cycle k, and the corresponding values, to the reference cycle n_ref, and second differences between mean values of said integral component for a minimum flow rate delivered by the first pump, to the current cycle k, and the corresponding values, to the reference cycle n_ref, are greater than the corresponding differences for the previous cycle k-1 or the antepenultimate cycle k-2.

In one embodiment, the device further comprises alerting means for alerting the driver of the detection of fouling of the fuel filter by the determination means.

These warning means include, for example, a visual element such as a light, or an audible alarm.

In one embodiment, the determining means is adapted to detect a stabilized operating phase when, during a time interval greater than a threshold time interval, the fuel temperature is between a minimum temperature and a maximum temperature, the variation of the pressure measured in the injection manifold is less than a pressure threshold, the variation of the rotation speed of the engine is less than a rotation speed threshold, and the variation of the fuel flow delivered by the first pump is below a flow threshold.

According to another aspect of the invention, there is also provided a method for detecting the fouling of a fuel filter of a fuel supply system of an internal combustion engine, particularly a motor vehicle, in a which regulates, by Proportional Integral and Derivative control, the fuel flow supplying the high pressure pump of a pump assembly comprising a first low pressure pump and a second high pressure pump arranged in series. Fouling of the fuel filter is determined from the integral component of said fuel flow control supplying the high pressure pump.

• la figure 1 est une vue schématique d'un mode de réalisation d'un système selon un aspect de l'invention ;
• la figure 2 représente une caractéristique dite croissante du débit en fonction de la commande d'un actuateur de débit ;
• la figure 3 représente une caractéristique dite décroissante du débit en fonction de la commande d'un actuateur de débit ;
• la figure 4 illustre la boucle de régulation choisie en fonction du point de fonctionnement du moteur ;
• la figure 5, représente un mode de réalisation du régulateur PID en structure parallèle ;
• la figure 6, illustre les composantes Proportionnelle, Intégrale et Dérivée, en fonction des phases de fonctionnement ;
• la figure 7, illustre la variation de la perte de charge aux bornes du filtre à carburant en fonction du débit de carburant, au cours du temps ;
• la figure 8 représente la perte de charge aux bornes du filtre à carburant en fonction du débit instantané à différents kilométrages du véhicule,
• les figures 9 et 10 illustrent un exemple de fonctionnement du procédé selon un mode de réalisation de l'invention, et
• la figure 1 1 représente un exemple de réalisation de l'invention comprenant une cartographie à deux dimensions.

Other objects, features and advantages of the invention will appear on reading the following description, given solely by way of non-limiting example, and with reference to the appended drawings, in which:

• the figure 1 is a schematic view of an embodiment of a system according to one aspect of the invention;
• the figure 2 represents a so-called increasing flow characteristic as a function of the control of a flow actuator;
• the figure 3 represents a so-called decreasing characteristic of the flow rate as a function of the control of a flow actuator;
• the figure 4 illustrates the control loop chosen according to the operating point of the motor;
• the figure 5 , represents an embodiment of the PID regulator in parallel structure;
• the figure 6 , illustrates the components Proportional, Integral and Derivative, depending on the phases of operation;
• the figure 7 , illustrates the variation of the pressure drop across the fuel filter as a function of the fuel flow, over time;
• the figure 8 represents the pressure drop at the fuel filter terminals as a function of the instantaneous flow at different mileage of the vehicle,
• the figures 9 and 10 illustrate an example of operation of the method according to one embodiment of the invention, and
• the figure 1 1 represents an exemplary embodiment of the invention comprising a two-dimensional map.

On the figure 1 is schematically shown a reference internal combustion diesel engine 1, supplied with fuel by a fuel supply system. The invention can also be applied to other types of direct or indirect injection engines high pressure fuel. In this example, the engine 1 comprises four cylinders, and the fuel supply system comprises four reference injectors 2, each connected by a high pressure conduit 3 to the common injection rail 4, also called "injection rail "and which constitutes a high-pressure accumulator for the fuel to be injected.

The fuel supply system comprises a low pressure booster pump 5a which draws fuel into the tank 6 of the vehicle via a low pressure circuit 7. The pump 5a, associated with a mechanical pressure regulator , not shown on the figure 1 , its function is to stabilize the pressure at the inlet of a pump 5b at high pressure. The two pumps 5a and 5b constitute what will be called in the rest of the description the pumping assembly 5a, 5b. The low-pressure booster pump 5a can be mechanically driven by being integrated with the high-pressure pump 5b, itself driven mechanically by the motor 1.

Alternatively, the booster pump 5a may be independent of the high pressure pump 5b and, for example, driven by an electric motor.

A flow actuator, or flow control valve 9, is arranged between the low pressure pump 5a and the high pressure pump 5b to adjust the amount of fuel supplied to the high pressure pump 5b, then the injection manifold 4 The feed system also comprises, between the outlet of the pump 5b and the injection manifold 4, on the duct 10, an optional pressure actuator, or optional controlled valve 11 for regulating the pressure of the pressure. fuel accumulated in the fuel rail 4.

In addition, the system includes a return circuit 12 for the delivery of fuel from the pump assembly 5a, 5b, injector ducts 3, and the discharge of the high pressure portion. On the figure 1 , the return circuit 12 is mounted in communication with the two valves 9 and 11 and with a single conduit 3. Of course, in reality, the return circuit 12 communicates with all the conduits 3. Finally, the engine 1 and its system The electronic control unit 13 comprises conventional components, such as microprocessors, hard EEPROM memories and RAM type buffers.

Furthermore, the electronic control unit 13 receives input information 14 via a connection 15. This information 14 comes from different sensors placed on the engine 1 and related systems, such as the fuel injection system or the fuel injection system. air supply system, providing, for example, an estimate of the injected fuel flow Qinj.

A fuel filter 16 is disposed between the fuel tank 6 and the pumping assembly 5a, 5b so as to filter the fuel conveyed by the pipe 7 to the low-pressure pump 5a, as well as the fuel conveyed by the fuel pump. return circuit 12 to the tank 6.

The electronic control unit 13 processes the data it receives as input to define or calculate control levels outputted so as to control the entire system. The control levels are sent to the various actuators that participate in the control of the ancillary systems and therefore the engine 1. More particularly, the control levels are transmitted via a connection 18 to the injectors 2, via a connection 19 to the first controlled valve 9 of regulating the fuel flow supplying the high-pressure pump 5b, and via a connection 20 to the second controlled valve 11 for regulating the fuel pressure accumulated in the common injection rail 4.

More precisely, the information 14 transmitted to the electronic control unit 13, such as the temperature of the engine coolant, the rotational speed of the engine, the temperature of the engine lubricating oil, the air pressure provided by a turbocharger, or the position of the accelerator pedal, are, for example, processed via functions or maps stored in an EEPROM type memory.

These maps make it possible to define the value of the desired pressure Pcons in the common injection ramp 4, or setpoint pressure. The control levels are then calculated according to this setpoint pressure value Pcons. Conventionally, the value of the pressure setpoint Pcons is compared with the value of the pressure actually measured Pmes in the fuel injection manifold 4. This measured value of the pressure Pmes is delivered to the electronic control unit 13 via a connection 21 connected to a pressure sensor 22 measuring the pressure in the injection manifold 4. A sensor 23 provides the electronic control unit 13 a measurement of the rotational speed R_motor of the engine 1 by a connection 24.

The rotation speed of the high-pressure pump 5b can be deduced from the speed of rotation of the engine 1, according to the mechanical drive ratio between the engine 1 and the high-pressure pump 5b. According to the pressure difference ΔP identified (ΔP = Pcons-Pmes), the electronic control unit 13 adjusts the control signals of the controlled flow control valve 9, and possibly the second controlled pressure control valve 11 so that the measured pressure Pmes reaches the pressure setpoint Pcons. If the pressure difference ΔP is positive, the electronic control unit acts to increase the flow rate and / or reduce the discharge or leakage. On the other hand, when the pressure difference ΔP is negative, the electronic control unit 13 acts in such a way as to reduce the flow rate and / or to increase the discharge. The controlled flow control valve 9 is controlled and controlled by regulation means 25 of the PID regulator type comprising components Proportional, Integral, and Derivative for continuously adjusting, in a closed loop, the control of the flow actuator 9. The electronic control unit 13 further comprises a module 25a for determining the fouling of the fuel filter 16, from the integral component of said regulating means 25, described in more detail later, and related to the pressure drop ΔP across the fuel filter 16.

Similarly, the second controlled pressure control valve 11 can be regulated by regulation means 26 of the PID regulator type. The electronic control unit 13 further comprises means 25 for determining the setpoint pressure Pcons in the injection manifold 4. This setpoint pressure Pcons is transmitted to the regulation means 25 via a connection 28, and control means 26 by a connection 29 derived from the connection 28.

If the flow actuator 9 has a so-called increasing flow characteristic as a function of the control, for example as shown in FIG. figure 2 when the pressure variation ΔP in the injection manifold 4 is positive (ie Pcons> Pmes), the regulating means 25 act to increase the control of the controlled flow control valve 9 in order to increase the quantity fuel entering the high-pressure part. In addition, when the pressure variation ΔP in the injection manifold 4 is negative (ie P _mes > P _cons ), the regulation means 25 act in such a way as to reduce controlling the controlled valve 9 to reduce the amount of fuel entering the high pressure portion.

On the other hand, if the controlled flow control valve 9 has a so-called decreasing flow rate characteristic as a function of the control, as illustrated in FIG. figure 3 the control of the controlled valve 9 is reversed. When the pressure difference ΔP in the injection manifold 4 is positive (ie Pcons> Pmes), the regulating means 25 act in such a way as to reduce the control of the controlled flow control valve 9 so as to increase the flow rate fuel entering the high pressure portion, and when the pressure difference ΔP is negative (ie Pmes> Pcons), the control means 25 increase the control of the controlled valve 9 flow control to reduce the flow rate of fuel entering the high pressure part.

The system of figure 1 comprises the optional controlled pressure control valve 11 which is generally closed-loop controlled with a PID regulator in a manner similar to that used for the controlled flow control valve 9. Depending on the operating point of the motor 1, it is possible to use either the closed control loop on the controlled flow control valve 9 or the closed control loop on the controlled pressure control valve 11.

The figure 4 illustrates an example of the control loop used, for such a system provided with a flow actuator 9 and a pressure actuator 11, depending on the operating point of the engine 1. The closed loop on the pressure actuator It is used in cold start, idle, or low fuel flow. Indeed, during such operations, the higher reactivity of the pressure actuator 11 is preponderant. L / 11 pressure actuator is also very useful when the engine is driven ("overrun" in English), or, in other words, when the injection of fuel into the engine is zero. In this case, the use of the pressure actuator 11 makes it possible to reduce the pressure of the fuel in the injection manifold 4 by ensuring the return of the excess fuel to the tank 6 via the return circuit 12, which can not be guaranteed with regulation only on the controlled valve 9 of flow control. In other cases, that is to say for most operating points of the engine 1, regulation is performed on the flow actuator 9 in order to provide only the quantity of fuel necessary to achieve the desired pressure. In the injection manifold 4. The flow control is more tempting than a pressure regulation, but minimizes the hydraulic power dissipated and thus improves the fuel efficiency of the engine 1.

dans laquelle :

• U(t) est le signal de commande en sortie du module de régulation P.I.D.

$$\mathrm{ϵ}\left(\mathrm{t}\right)=\mathrm{\Delta P}\left(\mathrm{t}\right)=\mathrm{Pcons}\left(\mathrm{t}\right)-\mathrm{Pmes}\left(\mathrm{t}\right)$$

K_P est le coefficient de gain d'action Proportionnelle,
K_I est le coefficient de gain d'action Intégrale,
K_D est le coefficient de gain d'action Dérivée.
De manière classique, on a les rotations suivantes : $${\mathrm{K}}_{\mathrm{I}}\mathrm{=}\frac{{\mathrm{K}}_{\mathrm{P}}}{{\mathrm{T}}_{\mathrm{I}}}$$

dans laquelle :

• T_I est une constante de temps d'action Intégrale, en s, et $${\mathrm{K}}_{\mathrm{D}}\mathrm{=}{\mathrm{K}}_{\mathrm{P}}\mathrm{\cdot }{\mathrm{T}}_{\mathrm{D}}$$
  

dans laquelle :

• T_D est une constante de temps d'action Dérivée, en s.

The control module 25 of the controlled valve 9 regulating the flow of fuel supplying the high-pressure pump 5b may comprise, for example, components Proportional Integral and Derivative parallel structure, as illustrated on the figure 5 . This type of PID control has a mathematical model represented by a second-order linear integro-differentiation equation which has the form: $$\mathrm{U}\left(\mathrm{t}\right)\mathrm{=}{\mathrm{K}}_{\mathrm{P}}\mathrm{\epsilon }\left(\mathrm{t}\right)\mathrm{+}{\mathrm{K}}_{\mathrm{I}}\underset{\mathrm{0}}{\overset{\mathrm{t}}{\mathrm{\int }}}\mathrm{\epsilon }\left(\mathrm{\tau }\right)\mathrm{d\tau }\mathrm{+}{\mathrm{K}}_{\mathrm{D}}\frac{\mathrm{d\epsilon }\left(\mathrm{t}\right)}{\mathrm{dt}}$$

in which :

• U (t) is the control signal at the output of the PID control module

$$\mathrm{\epsilon }\left(\mathrm{t}\right)=\mathrm{.DELTA.P}\left(\mathrm{t}\right)=\mathrm{Pcons}\left(\mathrm{t}\right)-\mathrm{Pmes}\left(\mathrm{t}\right)$$

K _P is the Proportional action gain coefficient,
K _I is the integral gain coefficient of action,
K _D is the Derivative Action Gain coefficient.
In a classical way, we have the following rotations: $${\mathrm{K}}_{\mathrm{I}}\mathrm{=}\frac{{\mathrm{K}}_{\mathrm{P}}}{{\mathrm{T}}_{\mathrm{I}}}$$

in which :

• T _I is an integral action time constant, in s, and $${\mathrm{K}}_{\mathrm{D}}\mathrm{=}{\mathrm{K}}_{\mathrm{P}}\mathrm{\cdot }{\mathrm{T}}_{\mathrm{D}}$$
  

in which :

• T _D is a Derived Action Time constant, in s.

dans laquelle :

• P(m) est le terme d'action Proportionnelle, avec : $$\mathrm{P}\left(\mathrm{m}\right)\mathrm{=}{\mathrm{K}}_{\mathrm{P}}\mathrm{\cdot }\mathrm{\epsilon }\left(\mathrm{m}\right)$$
  
• I(m) est le terme d'action Intégrale, pouvant être évalué par la méthode des rectangles : $$\mathrm{I}\left(\mathrm{m}\right)\mathrm{=}\mathrm{I}\left(\mathrm{m}\mathrm{-}\mathrm{I}\right)\mathrm{+}{\mathrm{K}}_{\mathrm{I}}\cdot \mathrm{T}\mathrm{\cdot }\mathrm{\epsilon }\left(\mathrm{m}\right)$$
  
• D(m) est le terme d'action Dérivée, pouvant être évalue par un calcul en deux points : $$\mathrm{D}\left(\mathrm{m}\right)\mathrm{=}\frac{{\mathrm{K}}_{\mathrm{D}}}{\mathrm{T}}\left[\mathrm{\epsilon }\left(\mathrm{m}\right)\mathrm{-}\mathrm{\epsilon }\left(\mathrm{m}\mathrm{-}\mathrm{I}\right)\right]$$
  
  et
• T est la période d'échantillonnage, en s, et n l'indice de l'échantillon.

By operating a discretization showing the three actions, Proportional, Integral and Derivative, we can ask the following equation: $$\mathrm{U}\left(\mathrm{m}\right)\mathrm{=}\mathrm{P}\left(\mathrm{m}\right)\mathrm{+}\mathrm{I}\left(\mathrm{m}\right)\mathrm{+}\mathrm{D}\left(\mathrm{m}\right)$$

in which :

• P (m) is the Proportional action term, with: $$\mathrm{P}\left(\mathrm{m}\right)\mathrm{=}{\mathrm{K}}_{\mathrm{P}}\mathrm{\cdot }\mathrm{\epsilon }\left(\mathrm{m}\right)$$
  
• I (m) is the term of integral action, which can be evaluated by the method of the rectangles: $$\mathrm{I}\left(\mathrm{m}\right)\mathrm{=}\mathrm{I}\left(\mathrm{m}\mathrm{-}\mathrm{I}\right)\mathrm{+}{\mathrm{K}}_{\mathrm{I}}\cdot \mathrm{T}\mathrm{\cdot }\mathrm{\epsilon }\left(\mathrm{m}\right)$$
  
• D (m) is the derivative action term, which can be evaluated by a two-point calculation: $$\mathrm{D}\left(\mathrm{m}\right)\mathrm{=}\frac{{\mathrm{K}}_{\mathrm{D}}}{\mathrm{T}}\left[\mathrm{\epsilon }\left(\mathrm{m}\right)\mathrm{-}\mathrm{\epsilon }\left(\mathrm{m}\mathrm{-}\mathrm{I}\right)\right]$$
  
  and
• T is the sampling period, in s, and n is the sample index.

The integral action 1 acts all the time and makes it possible to eliminate the permanent static error with respect to a setpoint or disturbance step, while the proportional actions P and derivative D act only in dynamic. The proportional action P makes it possible to increase the reaction speed with respect to a high setpoint gradient, whereas the derivative fraction D tends to stabilize the system by bringing a phase advance. This means that in stabilized operation, the proportional actions P and derivative D are zero, and that the control value is directly given by the integral action I (cf. figure 6 ).

dans laquelle:

• R_moteur est la vitesse de rotation du moteur, en tr/min ;
• Qinj est la consigne de débit de carburant à injecter en mg/coup, ou en m³/coup ;
• C_G,i est la commande globale de l'actuateur de débit 9 à l'instant i ;
• C_N(R_moteur ;Qinj) est la commande nominale de l'actuateur de débit, par exemple cartographiée en fonction du débit de carburant à injecter et de la vitesse de rotation du moteur ; et
• C_C(I_i) est la commande corrective à ajouter à la commande nominale de l'actuateur de débit qui dépend uniquement de faction intégrale du correcteur P.I.D en phase stabilisée.

Thus, in the stabilized phase, the corrective control added to the nominal control of the flow actuator 9 depends solely on the PID regulator integral portion 1 It is therefore possible to set, in the stabilized phase, for a given pair of values of the rotational speed of the engine and the fuel flow rate to be injected, for a given instant i, the following inequation: $${\mathrm{VS}}_{\mathrm{BOY\; WUT}\mathrm{,}\mathrm{i}}\mathrm{=}{\mathrm{VS}}_{\mathrm{NOT}}\left(\mathrm{R\_moteur}\mathrm{;}\mathrm{QINJ}\right)\mathrm{+}{\mathrm{VS}}_{\mathrm{VS}}\left({\mathrm{I}}_{\mathrm{i}}\right)$$

in which:

• R_motor is the rotational speed of the engine, in rpm;
• Qinj is the fuel flow instruction to inject in mg / stroke, or in m ³ / stroke;
• C _{G, i} is the global control of the flow actuator 9 at time i;
• C _N (R_moteur; Qinj) is the nominal control of the flow actuator, for example mapped as a function of the fuel flow to be injected and the speed of rotation of the engine; and
• C _C (I _i ) is the corrective command to be added to the nominal control of the flow actuator which depends solely on the integral action of the stabilized phase PID corrector.

If the system were perfect, the value of the integral action in the stabilized phase would be zero. But given the manufacturing dispersions and drifts over time (aging) components of the fuel system, operating deviations from the nominal system are inevitable. Also, the value of integral action in stabilized operation is the image of the corrections made to the fuel system in relation to the nominal operation.

Over time, the fuel filter 16 clogs or at least partially clogs, which can disrupt the proper functioning of the fuel supply system of the internal combustion engines by generating an increase in the pressure drop in the inlet of the pumping assembly. As illustrated on the figure 7 , the pressure drop ΔP__filter across the fuel filter 16 follows, in general, a second-degree law, depending on the fuel flow through it. The pressure drop ΔP_filtre across the fuel filter 16, for a given fuel flow, increases over time with the degree of fouling of the filter. When this pressure loss AP_filtre exceeds a critical threshold AP_filtre_critique, malfunctions of the engine 1 may appear.

It is then necessary to change the fuel filter 16. From an engine control point of view or control of the fuel supply, the increase in the pressure drop ΔP_fit the terminals of the fuel filter 16 generated by the increase the degree of clogging of the filter results in a couple of values of the rotational speed of the engine and the quantity of fuel to be injected, by a reduction of the fuel pressure in the injection manifold 4 with respect to the desired setpoint pressure (ie ΔP = Pcons - Pmes <0) in open loop. On the other hand, in a closed loop, this loss of pressure in the injection manifold 4 is naturally compensated by an increase in the value I of the integral action. This increase in I has the effect of increasing the quantity of fuel sent into the high-pressure part by increasing the corrective command C _C (I). In other words, the behavior of the value 1 of the integral term of the PID regulator in stabilized phase can be correlated with the pressure drop ΔP_filtre at the terminals of the fuel filter 16. We therefore have: $$\mathrm{I}\mathrm{=}\mathrm{f}\left(\mathrm{\Delta P\_filtre}\right)$$

At the critical pressure loss ΔP_critical_filter corresponds to an I_critical critical value of the integral component I.

The invention makes it possible, on the one hand, to follow the revolution of the fouling of the fuel filter 16 by not observing the integral term I in the stabilized phase of operation of the engine 1, in particular by means of the determination module 25a, and, on the other hand, to warn the driver that he must change his fuel filter 16, when the integral component reaches the I_critical threshold. The driver can be alerted by means of an alert module 30 connected to the electronic control unit 13 via a connection 31. The warning module 30 may for example comprise a light or a voice interface.

The evaluation of the degree of fouling of the fuel filter is based on the observation and analysis of the integral term I of the regulator P.I.D 25 in the stabilized phase.

A stabilized phase can be defined by variations ΔP_mes of the pressure Pmes measured in the injection manifold 4, variations ΔR_motor speed R_motor of rotation of the engine 1, and ΔQinj variations of the quantity Q _inj fuel to inject, not exceeding certain thresholds during a T_laps time interval. For example, a stabilized phase can be defined by the following system: $$\{\begin{array}{l}\mathrm{\Delta P\_mes}<\mathrm{threshold\_P\; with\; threshold\_P}=10\mathrm{bars}\\ \mathrm{\Delta R\_moteur}<\mathrm{threshold\_R\_motor\; with\; threshold\_R\_motor}=100\mathrm{tr}/\min \\ \mathrm{\Delta Qinj}<\mathrm{threshold\_Qinj\; with\; threshold\_Qinj}=2,5\mathrm{mg}/\mathrm{stroke}\\ \mathrm{T\_laps}>\mathrm{threshold\_T\_laps\; with\; threshold\_T\_laps}=5\mathrm{seconds}\end{array}$$

Then, we take into account that the integral term I is a corrective factor of the system. In other words, it corrects all the dispersions and drifts of the fuel system, but not the dispersions and drifts of a specific component of the fuel system. It is therefore necessary to define a criterion making it possible to take into account only the influence of the fouling of the fuel filter 16 on the term I. For this purpose, as the phenomenon of fouling of the fuel filter 16 is a phenomenon rapid in comparison with the wear or aging of any other component of the fuel supply system, an analysis of the evolution of the integral term I on relatively short observation cycles is carried out over the duration of use of a vehicle. For example, this analysis can be done at the full fuel level (tank filling) or thousand kilometers.

The integral term I is learned at each stabilized phase transition, around predetermined operating points ("breakpoints", in English) on the field of possible values of the engine rotation speed and the quantity of engine. fuel to be injected.

While the rotational speed of engine R_motor, in rev / min, and the setpoint of the quantity of fuel to be injected Qinj, in mg / stroke, are information classically available in an electronic control unit, the flow of fuel passing through the fuel filter 16 is not. The flow rate of fuel passing through the fuel filter 16, equal to the flow rate of fuel Q_pump passing through the low-pressure pump 5a, in liters / hour, can not be simply connected to the setpoint Qinj of the quantity of fuel to be injected into the engine 1. One solution, to dispose of this data Q_pump, is to equip the pump with a flowmeter electrically connected to the electronic control unit 13. However, for a given pump, it is easy to know the flow rate Q_pump from the speed of rotation of the pump R_pumpe and the amount of fuel to inject Qinj. As soon as the rotational speed of the pump R_pump is deduced from the rotational speed of the motor R_motor by a mechanical drive ratio a, it is then possible to work as a function of values of the speed of rotation of the pump R_pump and flow rate of the pump Q_pump instead of values of the speed of rotation of the motor R_motor and the flow rate to inject Qinj.

$$\mathrm{R\_pompe}\mathrm{=}\mathrm{\alpha R\_moteur}$$

Also, we obtain the following equations: $$\mathrm{Qpompe}\mathrm{=}\mathrm{boy\; Wut}\left(\mathrm{R\_pompe}\mathrm{;}\mathrm{QINJ}\right)$$

$$\mathrm{R\_pompe}\mathrm{=}\mathrm{\alpha R\_moteur}$$

At the end of each observation cycle k (detection of reservoir filling or thousand kilometers), the average evolution of the integral term learned I_app_moyen is determined on all the pairs of values (R_pump, Q_pump).

dans laquelle :

• n est le nombre de valeurs apprises au cours du cycle k pour un couple (R_pompe_j ; Q_pompe_l) donné ;
• i est la i^ème valeur apprise (i compris entre 1 et n) au cours du cycle k pour un couple (R_pompe_j ; Q_pompe_l) donné ;
• j est la j^ème valeur de la vitesse de rotation de la pompe à basse pression 5a parmi les vitesses de rotation de la pompe 5a fixées par les points de fonctionnement prédéfinis ; et
• 1 est la 1^ème valeur de débit de carburant de la pompe à basse pression 5a parmi les débits de pompe fixés par les points de fonctionnement prédéfinis.

Thus for each pair _(j R_pompe; Q_pompe _l) predetermined, it can store all the learned values I_app (R_pompe; Q_pompe) of the integral term in order to extract an arithmetic average at the end of each cycle. This average would then be expressed at the end of each observation cycle k according to the following equation: $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\mathrm{=}\frac{\underset{\mathrm{i}\mathrm{=}\mathrm{1}}{\overset{\mathrm{not}}{\Sigma }}\mathrm{I\_app}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{i}\mathrm{,}\mathrm{k}}}{\mathrm{not}}$$

in which :

• n is the number of values learned in cycle k for a given pair (R_pump _j ; Q_pump _l );
• i is the i ^th value learned (i between 1 and n) during the cycle k for a given torque (R_pump _j ; Q_pump _l );
• j is the j ^th value of the rotational speed of the low pressure pump 5a among the rotational speeds of the pump 5a fixed by the predefined operating points; and
• 1 is the 1 ^th fuel flow value of the low-pressure pump 5a among the pump flow rates determined by the predefined operating points.

This solution according to the equation 10 is expensive in use of the computer memory, one can alternatively use a sliding average. Thus for a given pair of values (R_pompe _j ; Q_pompe _l ), the principle of the sliding average is to store in memory the last average value of I, and to calculate, with the new value learned from I, the new average. For the ^nth learning of the integral term during the cycle k, the average value learned for a couple (R_pompe _j ; Q_pompe _l ) given will then be expressed according to the following equation: $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\mathrm{=}\frac{\left(\mathrm{not}\mathrm{-}\mathrm{1}\right)\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{i}\mathrm{,}\mathrm{k}}}{\mathrm{not}}\mathrm{+}\frac{\mathrm{I\_app}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}}{\mathrm{not}}$$

From the devolution of the integral component I, the fouling of the fuel filter 16 is deduced.

The integral term I represents the corrective control to be applied to the nominal control of the flow actuator 9 to compensate for the dispersions and drifts of the fuel supply system. It is as if the drifts and / or the dispersions of the flow actuator 9 were corrected with respect to its nominal characteristic (cf. figure 2 ). The drifts and dispersions of the flow actuator 9 have known maximum and minimum limits of functionality (see the maximum and minimum envelopes represented on FIG. figure 2 ). These limitations are known manufacturing data. Thus, for a given pair of values (Rpump _j ; Q_pump _l ), a first critical function threshold is reached when the corrective command C _C (I) applied to the flow actuator 9 reaches or exceeds the control difference between the command associated with the maximum permissible characteristic C _max (R_pumpe _j ; Q_pumpe _l ) and the command associated with the nominal characteristic C _N (R_pumpe _j ; Q_pump _l ).

ou lorsque : $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n}\mathrm{,}\mathrm{k}}\le \mathrm{I\_neg\_max}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{+}\mathrm{Offset\_neg}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)$$

dans lesquelles
Offset_pos(R_pompe_j ; Q__pompe_l) et
Offset_neg(R_pompe_j ;Q_pompe_l) sont les seuils paramétrables en fonction des couples de valeurs (R_pompe_j ;Q_pompe_l)) afin de s'affranchir des incertitudes de mesures.
Dans le mode de réalisation décrit, la caractéristique du débit en fonction de la commande de l'actuateur débit 9 est croissante (cf figure 2), et la perte de charge générée par le filtre à carburant 16 se traduit par une commande corrective positive, seule l'inéquation 1 est intéressante. Par défaut, l'inéquation 2 ne peut pas être réalisée à cause d'un phénomène d'encrassement du filtre à carburant 16. Par contre, en considérant une caractéristique décroissante du débit en fonction de la commande pour l'actuateur débit 9 (cf figure 3), le raisonnement serait inversé.Likewise, a second critical function threshold is reached when the corrective command C _C (I) applied to the flow actuator 9 reaches or exceeds the control difference between the command associated with the minimum tolerated characteristic C _min (R_pump _j Q_pump _l ) and the control associated with the nominal characteristic C _N (R_pump _j , Q_pump _l ). Sacron that the integral term represents the corrective command, by naming I_pos_max (R_pump, Q_pump) the value of I which corresponds to the first critical threshold C _max (R_pumpe _j ; Q_pumpe _l ) -C _N (R_pompe _j ; Q_pompe _l ), and I_neg_max (R_pumpe _j ; Q_pumpe _l ) the value of I which corresponds to the second critical threshold C _min (R_pump _j ; Q_pump _l ) -C _N (R_pumpe _j ; Q_pump _l ), we can consider that a critical threshold of functionality is reached when : $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\mathrm{\ge }\mathrm{I\_pos\_max}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{+}\mathrm{Offset\_pos}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)$$

or when: $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\le \mathrm{I\_neg\_max}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{+}\mathrm{Offset\_neg}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)$$

in which
Offset_pos (R_pumpe _j ; Q__pump _l ) and
Offset_neg (R_pumpe _j ; Q_pump _l ) are the thresholds that can be parameterized as a function of the pairs of values (R_pump _j ; Q_pump _l )) in order to overcome measurement uncertainties.
In the embodiment described, the characteristic of the flow rate as a function of the control of the flow actuator 9 is increasing (cf. figure 2 ), and the pressure drop generated by the fuel filter 16 results in positive corrective control, only the inequation 1 is interesting. By default, the inequation 2 can not be achieved because of a fouling of the fuel filter 16. On the other hand, considering a decreasing characteristic of the flow rate as a function of the control for the flow actuator 9 (cf. figure 3 ), the reasoning would be reversed.

Consequently, as the pressure drop generated by the fouling of the fuel filter 16 is compensated by a positive integral action 1 (increase of the control), it can be considered that a critical threshold of fouling of the fuel filter is susceptible to be reached when the maximum functionality limit of the flow actuator is reached, ie when: $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\mathrm{\ge }\mathrm{Offset\_pos}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{+}\mathrm{I\_pos\_max}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)$$

It must then be verified whether the crossing of the critical threshold is effectively linked to the fouling of the fuel filter 16. The manufacturing dispersions of the components of the fuel supply system may cause deviations from the nominal operation. Thus, the first observation cycles k, for example of the order of 3 to 5, serve essentially to correct the manufacturing dispersions of the components rather than the drifts. By Therefore, the first observation cycles are used to establish a reference from which the plausibility of the fouling of the fuel filter 16 is studied. If n_ref is the number of cycles necessary to learn the dispersions of the system, our reference is given by: $$\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{=}\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n\_ref}}$$

Therefore, for a given pair of values (R_pumpe _j ; Q_pumpe _l ) and an observation cycle k> n_ref, the influence of the pressure drop generated by the fouling of the filter can be followed by observation of the difference I_app_moyen (R _pompe _j; Q _pompe _{l) k} -I_app_ref (R_pompe _j; Q_pompe _l).

ou $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}\mathrm{-}\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{\ge }$$

$$\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}2}\mathrm{-}\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)$$

Thus, when the warning can be given, the hypothesis according to which the system is under the influence of a clogging of the fuel filter 16 is correlated when, for the pairs of values (R_pompe _j ; Q_pompe _l ) which have undergone a minimum number of apprentices n_app_min during each cycle k-2, k-1 and k (with k-2 ≥ n_ref), one checks on the one hand the following inequalities: $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}\mathrm{-}\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{\ge }\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}\mathrm{1}}\mathrm{-}\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\ell }\right)$$

or $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}\mathrm{-}\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)\mathrm{\ge }$$

$$\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}2}\mathrm{-}\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)$$

ou $$\left[\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\mathrm{-}\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\right]\mathrm{\ge }\left[\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}2}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\mathrm{-}\left(\frac{\mathrm{I\_app\_moyen}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right){}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}2}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\right]\mathrm{+}\mathrm{\tau au}\left({\mathrm{R\_pomp}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max\_Q\_pompe\_min}\right)\mathrm{]}$$

For a given speed of rotation R_pump _j of the low pressure pump 5a among the pairs of values (R_pumpe _j ; Q_pumpe _l ) which verify the inequalities 4 or 5, it is furthermore verified, for the minimum pump flow rates Q_pompe_min and maximum Q_pump_max identified, the following inequalities: $$\left[\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\mathrm{-}\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\right]\mathrm{\ge }\left[\left(\frac{\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}{\mathrm{)}}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}\mathrm{1}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\mathrm{-}\left(\frac{\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}{\mathrm{)}}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}\mathrm{1}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\right]\mathrm{+}\mathrm{\tau au}\left({\mathrm{R\_pomp}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max\_Q\_pompe\_min}\right)\mathrm{]}$$

or $$\left[\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\mathrm{-}\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\right]\mathrm{\ge }\left[\left(\frac{\mathrm{I\_app\_moyen}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}2}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\mathrm{-}\left(\frac{\mathrm{I\_app\_moyen}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right){}_{\mathrm{n\_app\_min}\mathrm{,}\mathrm{k}\mathrm{-}2}}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\mathrm{-}\frac{\mathrm{I\_app\_ref}\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_min}\right)}{\mathrm{Q\_pompe\_max}\mathrm{-}\mathrm{Q\_pompe\_min}}\right)\right]\mathrm{+}\mathrm{\tau au}\left({\mathrm{R\_pomp}}_{\mathrm{j}}\mathrm{;}\mathrm{Q\_pompe\_max\_Q\_pompe\_min}\right)\mathrm{]}$$

In fact, generally, the pressure drop of the fuel filter 16 as a function of the instantaneous flow rate at different milestones changes according to the network of linear curves present. figure 8 , with an increase in the slope over the kilometers.

The figures 9 and 10 illustrate an operationally free of the method according to the invention.

The method starts with an initialization of the indices k and i (step 40), respectively to the values zero and one. Then we test (step 41) if the index k is greater than or equal to the reference index n_ref increased by two, and if the inequation 1 is verified. If these conditions (step 41) are not realized, a test (step 42) is carried out, in a loop, if one is in a stabilized operating phase, and if, in addition, the fuel temperature is between a minimum temperature Tmin and a maximum temperature Tmax in order to avoid influences of the fuel temperature on the system, and in particular the phenomena of waxing and fuel viscosity variation.

If a stabilized operating phase is detected (step 42), then the integral term I_app (R_pump _j , Q_pumpe _l ) _{i, k} (step 43, passing through A) is stored, and the calculation and storage (step 44) of the mean value I_app_moyen (R_pompe _j , Q_pompe _l ) _{i, k} according to the equation 11. On increments of an index i (step 45) and one tests (step 46) if the current cycle k is completed.

If this is not the case, we return to step 41, and if it is the case, we go to the next cycle by incrementing the index k (step 47). The number of learnings i (R_pumpe _j ; Q_pumpe _l ) _{k is} memorized (step 48), and it is tested (step 49) whether the current index k is equal to the reference index n_ref.

If the index k is not equal to the reference index n_ref, it returns to step 41 (via B), otherwise, the integral reference term I_app_ref (R_pompe _j ; Q_pompe _l ) is stored according to Equation 12 (step 50), and return to step 41 (through B).

On the other hand, when the test conditions of the step 41 are carried out, the values of the pairs (R_pompe _j ; Q_pompe _l ) which determine the following system of inequalities are determined (step 51): $$\mathrm{i}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)}_{\mathrm{k}}\mathrm{\ge }\mathrm{n\_app\_min\; and\; i}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)}_{\mathrm{k}\mathrm{-}\mathrm{1}}\mathrm{\ge }\mathrm{n\_app\_min\; or\; i}{\left({\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}\right)}_{\mathrm{k}\mathrm{-}\mathrm{2}}\mathrm{\ge }\mathrm{n\_app\_min}$$

Next, the minimum flow rates Q_pump_min and maximum Q_pump_max are determined (step 52) on the set of pairs of values (R_pumpe _j ; Q_pump _l ) which satisfy the inequation system 8.

We then test (step 53) whether the inequalities 4 or 5 and 6 or 7 are verified, and if this is not the case, go to step 42.

On the other hand, if the test conditions of step 53 are verified, the driver (step 54) is alerted by the warning means 30, that the fuel filter 16 is fouled. Thus, the driver can go to change the fuel filter and thus avoid operating impairments that can lead to vehicle breakdowns.

avec : $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n}\mathrm{,}\mathrm{k}}\mathrm{=}\mathrm{I\_app\_mémo}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{k}\mathrm{-}\mathrm{1}}\mathrm{+}\mathrm{\Delta I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{n}\mathrm{,}\mathrm{k}}$$

dans lequel :

• I_app_mémo est la dernière valeur de I apprise pour le couple (R_pompe_j;Q_pompe_l) au cycle précédent k-1.

The figure 11 describes an exemplary embodiment of the invention comprising a two-dimensional cartography dependent on the rotational speed of the engine and the target fuel flow rate to be injected into the engine, which is determined during the operation of the engine from the training and The advantage of this embodiment is the improvement over time of the reactivity of the system by direct compensation of the drifts of the system. The integral part I thus has only small deviations to correct, and consequently, with this type of adaptive corrections, during a cycle k, it is not the integral integral term I that is observed, but a ΔI variation of this global integral term. In this case, for a given pair (R_pompe _j ; Q_pompe _l ), equation 11 becomes the following system of equations 13: $$\mathrm{\Delta I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\mathrm{=}\frac{\left(\mathrm{not}-1\right)\mathrm{\Delta I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{-}\mathrm{l}\mathrm{,}\mathrm{k}}}{\mathrm{not}}\mathrm{+}\frac{\mathrm{\Delta I\_app\_}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}}{\mathrm{not}}$$

with: $$\mathrm{I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}\mathrm{=}\mathrm{I\_app\_mémo}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{k}\mathrm{-}\mathrm{1}}\mathrm{+}\mathrm{\Delta I\_app\_moyen}\mathrm{(}{\mathrm{R\_pompe}}_{\mathrm{j}}\mathrm{;}{\mathrm{Q\_pompe}}_{\mathrm{l}}{\mathrm{)}}_{\mathrm{not}\mathrm{,}\mathrm{k}}$$

in which :

• I_app_memo is the last value of I learned for the pair (R_pump _j ; Q_pump _l ) in the previous cycle k-1.

The present invention thus makes it possible, at reduced cost, to detect a fouling of the fuel filter with improved accuracy, and to warn the driver of a fouling rate necessitating the change of the fuel filter.

Claims (10)

Device for detecting fouling of a fuel filter (16) of a fuel supply system of an internal combustion engine (1), in particular of a motor vehicle, the fuel supply system comprising a pump assembly, comprising a first low pressure pump (5a) and a second high pressure pump (5b) arranged in series, and being arranged between a fuel injection ramp (4) provided with a pressure sensor ( 22) and a fuel tank (6), a first controlled valve (9) for regulating the fuel flow supplying the high pressure pump (5b), means (27) for determining a set pressure in the ramp injection device (4), and control means (25) of said first controlled valve (9) comprising Proportional Integral and Derivative components and a feedback loop for controlling the pressure measured by the sensor (22) on said pressure setpoint, the filter fuel nozzle (16) being arranged between the fuel tank (6) and the pumping assembly (5a, 5b), characterized in that it comprises means (25a) for determining the fouling of the fuel filter ( 16) from said integral component of said regulating means (25). An apparatus according to claim 1, wherein said determining means (25a) is adapted to evaluate the integral component of the regulating means (25) for stabilized operating phases. Apparatus according to claim 2, wherein said determining means (25a) is adapted to perform said evaluation cyclically, a cycle corresponding, for example, to the consumption of a threshold fuel quantity or to the course of a distance. threshold by the vehicle. Apparatus according to claim 2 or 3, wherein the determining means (25a) is adapted to detect fouling of the fuel filter (16) when a set of average values of said Integral component, dependent on the speed of rotation and the flow rate delivered by the first low pressure pump (5a), to the current cycle k, are greater than respective maximum thresholds depending on the speed of rotation and the flow rate delivered by the first low pressure pump (5b), the current cycle k being greater than or equal to a reference cycle number n_ref increased by at least two. Apparatus according to claim 4, wherein the determining means (25a) is adapted to use arithmetic or sliding averages. Device according to claim 4 or 5, wherein the determining means (25a) are adapted to detect a clogging of the fuel filter (16) when, in addition, the differences between said average values of said integral component, the current cycle k , and corresponding values, for the reference cycle n_ref, dependent on the rotational speed and the flow rate delivered by the first pump (5a), are greater than the corresponding differences for the previous cycle k-1 or the ante-penultimate cycle k- 2. Device according to Claim 6, in which the determination means (25a) are suitable for detecting clogging of the fuel filter when, in addition, differences between first differences between mean values of said integral component for a maximum delivered flow rate are provided. by the first pump (5a), to the current cycle k, and the corresponding values, to the reference cycle n_ref, and second differences between mean values of said integral component for a flow rate the minimum value delivered by the first pump (5a), to the current cycle k, and the corresponding values, to the reference cycle n_ref, are greater than the corresponding differences for the previous cycle k-1 or the antepenultimate cycle k-2. Device according to one of claims 4 to 7, further comprising warning means (30) for alerting the driver of the detection of fouling of the fuel filter (16) by the determining means (25a). ). Device according to one of claims 4 to 8, wherein the determining means (25a) are adapted to detect a stabilized operating phase when, during a time interval greater than a threshold time interval, the fuel temperature is between a minimum temperature and a maximum temperature, the variation of the pressure measured in the injection manifold (4) is less than a pressure threshold, the variation of the rotation speed of the engine (1) is less than a threshold rotational speed, and the variation of the fuel flow delivered by the first pump (5a) is less than a flow threshold. Method for detecting the fouling of a fuel filter (16) of a fuel supply system of an internal combustion engine (1), in particular of a motor vehicle, in which the proportional integral regulation is regulated and Derivative, the fuel flow supplying the high pressure pump (5b) of a pump assembly comprising a first low pressure pump (5a) and a second high pressure pump (5b) arranged in series, characterized in that the fouling of the fuel filter (16) is determined from the integral component of said fuel flow control supplying the high pressure pump (5b).

Images