METHOD FOR VALIDATING A DIAGNOSIS OF RICHNESS DEVIATIONS UNDER SIMILAR CONDITIONS

Description

[0001] The field of the invention relates to a method for validating a fuel mixture deviation diagnosis implemented by a control unit of a heat engine.

[0002] Internal combustion engine actuators in vehicles can exhibit behavior that differs from the behavior models integrated into the engine control unit (ECU) due to manufacturing variations, wear, fouling, and the quality of the actuator modeling. In the case of actuator models related to the intake and injection ports, this discrepancy can lead to fuel mixture deviations. This situation results in increased fuel consumption and / or higher emissions. It can also negatively impact driving experience. It is necessary to correct these mixture deviations throughout the vehicle's lifespan.

[0003] Typically, this correction is performed by engine control models, which implement the fuel mixture regulation function. This function continuously adjusts the injector timing based on the fuel mixture measurement provided by the exhaust gas mixture sensor. Furthermore, to optimize this correction, the engine control unit (ECU) simultaneously executes an engine control learning function responsible for memorizing the necessary correction for specific engine operating conditions. More precisely, at a stabilized point within the engine operating conditions, as soon as the learning conditions are met, the learning function memorizes the correction value determined by the engine control model responsible for fuel mixture regulation. For example, the learning process can be implemented using a self-adaptive neural network.Simultaneously, the control unit communicates the value being stored, which causes the fuel mixture correction value to converge. This minimizes the regulation workload and improves the correction dynamics.

[0004] The learning function plays a major role when the fuel mixture control model cannot correct the mixture, particularly when the mixture sensor is unavailable, and significantly improves engine control during load transients by minimizing the workload of the fuel mixture control function. Document FR3057031A1, describing a mixture correction method filed by the applicant, is known. The described method is a least-squares minimization optimization technique based on analyzing the errors present in the system to determine terms of the correction function, with the aim of correcting the errors by improving the models. Document FR2979390A1, describing an optimization technique based on a Kalman filter, is also known.

[0005] Furthermore, the prior art is known from document US10612484B2.

[0006] Regulations concerning on-board diagnostics require manufacturers to monitor learning functions to alert the driver in the event of a fault that could lead to exceeding acceptable emission levels. Typically, fuel mixture deviation diagnostics involve checking the values ​​taken by the adaptive correction systems against limit values. Furthermore, some regulations require supplementing diagnostics with a fault resolution validation mechanism under similar conditions, commonly referred to as "Similar Conditions."

[0007] We illustrate by figure 1 The principle of this validation mechanism, as implemented in the prior art for diagnosing a fuel mixture control learning function, is illustrated. Curves C1 to C5 are represented on a time axis during which two learning phases, A1 and A2, occur. These phases determine, respectively, an adaptive correction that will be applied by the fuel mixture control function. The upper part of curve C1 represents the number of fuel mixture measurements at a stabilized engine operating point. Each learning phase, A1 and A2, includes the calculation of a series of temporary corrections—intermediate values ​​not applied by the fuel mixture control function—which are calculated from the fuel mixture deviation measurements. The value of each adaptive correction resulting from a learning phase is illustrated by curve C3.Curve C4 represents the two time points t1 and t2 when a fuel mixture deviation diagnosis is performed. Each diagnosis takes into account the series of temporary corrections used to detect a fault. Curve C5 represents the fuel mixture deviation values ​​during the learning phases.

[0008] The principle for validating the end of a fault under similar conditions is as follows. If a fault is detected for the first adaptive setting at time t1, the diagnostic system stores the instantaneous engine parameters, which are the parameters of the final adaptive setting applied by the control function. The fault detection triggers an alert to the driver and the selection of these instantaneous engine parameters to identify the reference engine operating point SC used for the validation mechanism under similar conditions.Then, at the end of the A2 learning phase, if a fault termination is detected for the second adaptive system at time t2, and provided that a reference point has been previously stored, the diagnostic system is authorized to also validate the fault termination for the first adaptive system if the test performed under similar conditions to the reference point is positive. Specifically, this test is positive if the differences between the engine parameters of the reference point SC and the engine parameters of the second adaptive system are less than a given threshold. A positive result also clears the driver alert that was generated for the first adaptive system.

[0009] This mechanism has drawbacks in terms of false alarms. Indeed, it sometimes happens that a fault triggered by a temporary fix ultimately has no impact on the system or pollutant emissions at the moment the final adaptive adjustment is applied to the actuator models. This is because the engine operating point for the final adaptive adjustment may differ from the engine operating point that triggered the alert.

[0010] Therefore, there is a need to address the aforementioned problems. Furthermore, one objective of the invention is to reduce the persistence of false alerts regarding fuel mixture deviation and to improve the validation mechanism under similar conditions for fuel mixture regulation functions using learning functions.

[0011] More specifically, the invention relates to a method for validating a fuel mixture deviation diagnosis implemented by a control unit of an internal combustion engine, comprising the following steps: The control of a richness regulation function applying an adaptive correction, where each adaptive correction is determined at the end of a learning phase, each learning phase succeeding and including the calculation of a series of temporary corrections calculated from measurements of the richness deviation, The richness deviation diagnosis consisting of: o In case of detection of a failure for a first adaptive applied by the regulation function, to memorize a reference engine operating point, o In case of detection of an end of failure for a second adaptive, to validate the end of failure of the first adaptive according to a test carried out under criteria of similar conditions with said reference point.

[0012] According to the invention, the method further comprises selecting said reference point of the first adaptive from the series of temporary corrections according to the value of the richness deviation of each correction, and the test under similar conditions criterion is carried out by comparing each temporary correction of the second adaptive with the reference point, this test comprising, for each temporary correction of the second adaptive, the calculation of the difference between the reference point and the operating point of each temporary correction, and the comparison of said difference with respect to a predetermined threshold.

[0013] According to one variant, the memorization of the reference point consists of recording all or part of the instantaneous parameters of the engine including engine speed, engine load and engine water temperature.

[0014] More specifically, according to this last variant, the calculation of the operating point difference consists of calculating the difference in engine speed and / or engine load and / or water temperature.

[0015] According to one variant, a failure is detected when at least one temporary fix in a series of fixes has a richness deviation greater than a predetermined limit.

[0016] According to a first method of selecting the reference point, the process further includes recording the operating point of each temporary fix during the learning phase, and the selection of said reference point consists of selecting the operating point of the temporary fix exhibiting the highest richness deviation among said temporary fixes.

[0017] According to a second method of selecting the reference point: During the learning phase for each temporary fix, the comparison of the richness deviation values ​​of the current fix and the previous fix of said series, and the recording of the operating point of the fix with the highest richness deviation in a temporary variable, At the end of said learning phase, in case of detection of a failure, the operating point of said temporary variable is selected for said reference point.

[0018] According to a third selection method, the process further includes recording the operating point of each temporary fix during the learning phase, calculating the average deviation value for the series of temporary fixes of said learning phase, and in that the selection of said reference point consists of selecting the operating point of the temporary fix exhibiting the deviation furthest from the average value.

[0019] The invention also provides for a thermal engine control unit comprising a mixture regulation module and a mixture deviation diagnostic module, said unit being configured to implement the diagnostic validation method according to any one of the preceding embodiments.

[0020] The invention provides for an automobile comprising a thermal engine controlled by said control unit.

[0021] Other features and advantages of the present invention will become more apparent upon reading the following detailed description, which includes embodiments of the invention given by way of non-limiting examples and illustrated by the accompanying drawings, in which: [ Fig.1 ] represents the implementation of the validation mechanism as described in the preamble describing the prior art; [ Fig.2 ] is a flowchart representing a first embodiment of the validation process according to the invention for the phase of memorizing the reference point used by the test under similar conditions criteria; [ Fig.3 ] is a flowchart representing the continuation of the sequence of the first embodiment of the validation process according to the invention, specifically during testing under similar conditions; [ Fig.4 ] represents graphs illustrating the mechanism for validating the end of a failure according to the method according to the invention.

[0022] The invention finds application in internal combustion engine actuator control models, and in particular in the mechanism for validating the end of a fault in a fuel mixture deviation warning diagnostic under similar conditions. The invention is described for a fuel mixture regulation function implementing a learning function used to drive actuator control models.

[0023] In this description, the learning function implements a learning phase during which temporary correction values ​​converge to a final value in order to reduce the fuel mixture deviation. The learning function is adapted to memorize adaptive fuel mixture corrections, based, for example, on a matrix of engine speed and load variables. The fuel mixture deviation value is estimated from measurements of the oxygen content of the engine exhaust gas, notably by the exhaust oxygen sensor.

[0024] The term "temporary correction" refers to values ​​calculated internally by the internal combustion engine control unit that are not applied by the actuator models. To determine each adaptive correction, a learning phase involves calculating a series of temporary corrections until the learning function determines that the correction value reaches a sufficient level of confidence with respect to one or more criteria, such as a minimum number of measurements, a convergence criterion, or a confidence index.

[0025] The term "adaptive correction" refers to the final value resulting from the convergence of the learning phase. This value is stored and retrieved by the learning function to control the actuator models when the engine enters a specific operating range, particularly in terms of speed and load, associated with the adaptive correction. The adaptive corrections adjust the actuators of the intake and / or injection branches of the internal combustion engine to control the fuel mixture, specifically the intake valve opening and / or the fuel injection duration.

[0026] The diagnostic validation process is implemented by a control unit comprising a fuel mixture control module and a fuel mixture deviation diagnostic module. The control unit includes means for calculating a series of temporary corrections from fuel mixture deviation measurements and determining a final adaptive correction to be applied by the control function. The control unit further includes fuel mixture deviation diagnostic means capable, in the event of a fault detected for a first adaptive correction applied by the control function, of storing a reference engine operating point, and in the event of a fault termination for a second adaptive correction, of validating the fault termination of the first adaptive correction based on a test performed under conditions similar to said reference point.

[0027] The control unit is equipped with an integrated circuit computer and electronic memory, the computer and memory being configured to execute the process according to the invention. However, this is not mandatory. Indeed, the computer could be external to the engine control unit, while still being coupled to it. In this latter case, it could itself be configured as a dedicated computer including, for example, a dedicated program. Consequently, the control unit, according to the invention, can be implemented in the form of software modules, electronic circuits, or hardware, or a combination of electronic circuits and software modules.

[0028] There figure 2 and the figure 3 represent a first embodiment of the process according to the invention in the form of a flowchart illustrating the algorithm of the sequence of validation of an end of failure under criteria of similar conditions.

[0029] More specifically, with reference to the figure 2 The process includes an initial learning phase A1, which performs loops to calculate temporary fuel mixture corrections at stabilized operating points under engine speed and load conditions. figure 4 illustrates the same sequence and the same graphs C1 to C5 as those of the figure 1 The difference is that this time it corresponds to the application of the method according to the invention. Each loop includes a measurement E20 of the air-fuel ratio, illustrated by curve C1 (number of measurements) and curve C5 (value of air-fuel ratio deviation), for example by measuring the exhaust oxygen sensors, then the calculation E21 of a temporary correction from the air-fuel ratio deviation measurements, illustrated by curve C2. Each temporary correction is stored in the memory of the control unit.

[0030] In each iteration, the process includes an end-of-training evaluation step E22. The training function checks whether the correction value reaches a sufficient confidence level with respect to one or more criteria, such as a minimum number of measurements, a convergence criterion, or a confidence index. If the confidence level is not reached, then steps E20, E21, and E22 are executed again.

[0031] At the end of the learning phase, at time t1, the process includes a step that controls the fuel mixture regulation function by applying an adaptive correction. This adaptive correction has the last value of the learning phase, as illustrated by curve C3. The process also includes a diagnostic step, E24, for fuel mixture deviation at time t1. This step consists of verifying the values ​​of all the temporary corrections taken during the learning phase against a predetermined threshold corresponding to a maximum deviation limit. At step E25, if at least one of the temporary corrections exceeds the predetermined threshold, the diagnosis is positive and an alert is generated for the driver.

[0032] Furthermore, for the purposes of the failure validation mechanism under similar conditions, the method according to the invention will select as the reference engine operating point SC that corresponding to the temporary correction that exhibited the largest fuel mixture deviation. To this end, the method includes a step E26 comparing the fuel mixture deviation value between each temporary correction, these values ​​being illustrated by curve C5 with reference to the figure 4 and a step to select the SC reference point, including all or part of the engine parameters for speed, load, and water temperature. During this step E26, the SC reference point is selected based on the mixture deviation values ​​of all temporary fixes so as to record the temporary fix with the largest mixture deviation, illustrated by the arrows on curves C2 and C5. Once the SC reference point recording E27 is performed, the validation process proceeds to a new learning phase.

[0033] Next, with reference to the figure 3 and the figure 4 , we illustrate the process for a second learning phase A2 following which we detect an end of richness deviation failure and for which we apply the validation mechanism under criteria of similar conditions using the reference point SC memorized during a previous sequence.

[0034] At step E28, the process initiates the second learning phase A2, during which loops are again performed to calculate temporary mixture corrections at stabilized operating points under engine speed and load conditions. Each loop comprises steps E28 for measuring the mixture, E29 for calculating a temporary correction, and E30 for evaluating the end of the learning process. These steps are identical to steps E20, E21, and E22 described in figure 2 Then, the process again includes an E31 step of controlling the fuel mixture regulation function by an adaptive correction, the last value of the learning phase at time t2 illustrated by curve C3 in figure 4 The process further includes a diagnostic step for the fuel mixture deviation E32 at time t2, consisting of verifying the values ​​taken by all the temporary corrections during the second learning phase A2 against the predetermined threshold. For this sequence, the diagnostic 32 detects the end of the failure for the temporary corrections calculated during phase A2. That is to say, the fuel mixture deviation value of all the temporary corrections, curve C5, is below the predetermined threshold.

[0035] Next, according to the invention, the method includes a verification step E33 to check that a reference point SC has been previously stored. If the result is positive, the method includes a test step E34 in which each temporary fix calculated during the second learning phase A2 is compared with the reference point SC to perform the test under similar conditions. More specifically, test E34 consists of calculating the difference between the operating point of each temporary fix and the reference point SC. Preferably, this step consists of calculating the difference in engine speed and / or engine load and / or water temperature. When this difference is less than a predetermined threshold for engine speed, load, and / or water temperature, the method according to the invention includes a validation step E35 at time t2 of the end of the fault previously resolved during the sequence of the figure 2 .

[0036] The diagnostic validation process further includes a second embodiment in which the SC reference point is detected before the diagnostic is performed. This differs from the first embodiment in that it does not require recording the engine parameters of each temporary correction until the diagnostic is executed. To this end, the process uses a temporary variable that, at each calculation loop of a temporary correction, records the operating point of the correction exhibiting the largest fuel mixture deviation between the current correction and the correction of the previous loop.

[0037] This second embodiment of the process is distinguished by the fact that, at the beginning of a learning phase, it initializes the temporary variable to a zero value, which is then updated in each calculation loop of the learning phase with the operating point of the major deviation correction. During the learning phase, the process includes, in each calculation loop, a comparison of the richness deviation values ​​of the current correction and the previous correction in the series, and the recording of the operating point of the correction with the highest richness deviation in the temporary variable.

[0038] Then, at the end of the learning phase, at time t1 with reference to the figure 4 If a fault is detected, the operating point stored in the temporary variable is selected for the reference point SC. If the diagnosis is negative, the temporary variable is reset to zero.

[0039] This principle can also be applied during the second learning phase A2 for the validation mechanism. Assuming that a reference point SC was stored during a previous diagnosis, in each loop of calculating a temporary fix, the process includes a step comparing the reference operating point SC with the operating point of the current temporary fix. More precisely, the process performs the test under similar conditions identically to the first embodiment, i.e., by calculating the differences in operating conditions, load, and / or water temperature and comparing the difference(s) with predetermined thresholds. If the test result is positive, the fault-ending diagnosis is temporarily validated. However, this validation is not yet effective during the learning phase.Then, during the diagnostic execution at the end of the learning phase, if an end of failure is detected, the process validates the test under criteria of similar conditions and cancels the alert.

[0040] A third embodiment is also envisaged in which the selection of the reference operating point SC is performed by selecting the temporary correction exhibiting the richness deviation furthest from the average of all the corrections in a learning phase. More precisely, with reference to the sequence of the figure 4The process includes recording the operating point of each temporary fix during the learning phase A1 and calculating the average value of the richness deviation for the series of temporary fixes in phase A1. During the execution of the diagnostic, at time t1 the selection of said reference point SC consists of selecting the operating point of the temporary fix exhibiting the richness deviation most offset from the average value, in values ​​lower or higher than the average value of richness deviation.

[0041] The process makes it possible to comply with regulations imposing a failure end validation mechanism under criteria of similar conditions for richness deviation diagnostics, and avoids the maintenance of false alerts when the point of realization of the diagnosis is not systematically correlated with the point of appearance of the failures.

Claims

1. Method for validating a diagnosis of air-fuel ratio deviation implemented by a control unit of an internal combustion engine, comprising the following steps: - Controlling an air-fuel ratio regulation function (E23, E31) applying a correction adaptive value, each correction adaptive value being determined at the end of a learning phase (A1, A2), each learning phase succeeding one another and comprising calculating a series of temporary correction values calculated from measurements of the air-fuel ratio deviation (E20, E28), - The air-fuel ratio deviation diagnosis consisting of: o In the event of detection of a fault (E25) for a first adaptive value applied by the regulation function, storing (E27) a reference engine operating point (SC), o In the event of detection of an end of fault (E32) for a second adaptive value, validating the end of fault (E35) of the first adaptive value based on a test (E34) performed under a similar conditions criterion with said reference point (SC), - Method being characterized in that it further comprises selecting (E26) said reference point (SC) of the first adaptive value from the series of temporary correction values depending on the air-fuel ratio deviation value of each correction value, and in that the test (E34) is performed by comparing each temporary correction value of the second adaptive value with the reference point (SC), the test comprising calculating the difference between the reference point (SC) and the operating point of each temporary correction value and comparing said difference with a predetermined threshold.

2. Method according to claim 1, wherein storing (E27) the reference point (SC) consists of recording all or part of the instantaneous engine parameters including engine speed, engine load and engine coolant temperature.

3. Method according to claims 1 and 2, wherein calculating the operating point difference comprises calculating the difference in engine speed and / or engine load and / or coolant temperature.

4. Method according to any one of claims 1 to 3, wherein a fault is detected when at least one temporary correction value exhibits an air-fuel ratio deviation greater than a predetermined limit.

5. Method according to any one of claims 1 to 4, further comprising storing the operating point of each temporary correction value during the learning phase (A1), and selecting the operating point of the temporary correction value exhibiting the highest air-fuel ratio deviation.

6. Method according to any one of claims 1 to 4, further comprising: - During the learning phase (A1), for each temporary correction value, comparing the deviation values of the current correction value and the previous correction value of the series, and storing the operating point of the correction value having the highest deviation in a temporary variable, - At the end of the learning phase, if a fault (E25) is detected, selecting the operating point stored in said temporary variable as the reference point (SC).

7. Method according to any one of claims 1 to 4, further comprising storing the operating point of each temporary correction value during the learning phase (A1), calculating the mean deviation value for the series of temporary correction values, and selecting the correction value whose deviation is the farthest from the mean value.

8. Control unit for an internal combustion engine comprising an air-fuel ratio regulation module and an air-fuel ratio deviation diagnostic module configured to implement the diagnostic validation method according to any one of claims 1 to 7.

9. Motor vehicle comprising an internal combustion engine controlled by a control unit according to claim 8.