Detecting backfire in a gaseous fuel internal combustion engine

Abstract

A method of detecting backfire in a gaseous fuel internal combustion engine is presented. The engine having a plurality of cylinders with respective intake ports connected to an air intake manifold, and one fuel injector per cylinder. The method comprises: monitoring intake manifold pressure and engine rotation; determining a plurality of synchronized intake manifold pressure values over corresponding engine segments, where an engine segment represents a predefined angular range corresponding to a given cylinder; determining the occurrence of a backfire condition in a respective cylinder by comparing one or more of said synchronized pressure values to at least one threshold (TH1, TH2, TH3, TH4).

Description

[0001] P-DELPHI-440 / WO

[0002] DETECTING BACKFIRE IN A GASEOUS FUEL INTERNAL COMBUSTION ENGINE

[0003] Technical field

[0004] The present invention generally relates to the operation of internal combustion engines running on gaseous fuel such as hydrogen. The invention more specifically relates to the detection of backfire in such engine.

[0005] Background Art

[0006] In the internal combustion engine, the ignition / combustion process is supposed to happen while the valves are closed, at the firing top dead center, inside the cylinder. One possible defect of the combustion process can be that for some reason the combustion occurs before full the intake valve(s) reaches the fully closed position, i.e. the air-fuel mixture is ignited in the intake manifold. This phenomenon is typically called backfire (or intake backfire).

[0007] Since the intake components of the engine are generally much weaker than the cylinder or exhaust system, the risk of engine damage during a backfire is very high. Therefore, it is crucial to detect any occurrence of backfire promptly.

[0008] On gaseous fuel engines, and especially hydrogen engines with port fuel injection (PFI) configuration (i.e. where fuel is injected before the intake valves), the risk of occurrence of backfire is by far higher than on other engines due to its high flammability.

[0009] Known backfire detection methods rely on temperature or pressure measurements. The prior art methods based on temperature are very slow to perform a detection, requiring a few seconds. The detection based on pressure is time based, and focused on single point injection systems; i.e. not capable to discriminate individual faulty injector or cylinder.

[0010] A further prior art backfire detection method is uses a vibration sensor, which is not present on standard mechanization, and the corresponding algorithm is incapable of issuing a robust counting information per cylinder. P-DELPHI-440 / WO

[0011] It is worth noting that in typical gas-fueled engines, such as those running on CNG, backfires are relatively rare, and engine designers have not historically considered backfire detection to be essential.

[0012] Technical problem

[0013] There is thus a need for an improved method for detecting engine backfire, that is particularly suited for hydrogen combustion engines.

[0014] General Description of the Invention

[0015] The present invention relates to a method for detecting backfire in a gaseous fuel internal combustion engine as claimed in claim 1.

[0016] The gaseous fuel internal combustion engine has a plurality cylinders with respective intake ports connected to an air intake manifold. One fuel injector (at least) is provided per cylinder. The fuel injector may be arranged for port fuel injection or for direct fuel injection.

[0017] The method comprises the steps of:

[0018] - monitoring intake manifold pressure and engine rotation;

[0019] - determining a plurality of synchronized intake manifold pressure values over corresponding engine segments, where an engine segment represents a predefined angular range corresponding to a given cylinder;

[0020] - determining the occurrence of a backfire condition by comparing one or more of said synchronized intake manifold pressure values to at least one threshold.

[0021] The inventive method is designed to detect backfire events in an engine, based on the intake manifold pressure. More specifically, synchronized intake manifold pressure values are determined; hence the pressure information is related to the corresponding crank angle. It is thus possible to relate the pressure information to a corresponding combustion segment of a cylinder.

[0022] In principle, any method for determining synchronized intake manifold pressure values may be used. P-DELPHI-440 / WO

[0023] Advantageously, each synchronized intake manifold pressure value is computed based on the intake manifold pressure data measured over a predetermined crank angle interval / range within the segment, in particular as an average of these measured pressure values. That is a synchronized intake manifold pressure value corresponds to a predetermined crank angle range.

[0024] In embodiments, the engine segments are combustion segments corresponding to respective cylinders. Typically, in a 4-stroke engine with N cylinders, a combustion segments has a width of 720 / N degrees. Each combustion segment is subdivided into subsegments defined between predetermined angular positions, preferably equally distributed between the lower and upper bounds of the combustion segment.

[0025] The intake manifold pressure is monitored continuously and several pressure values are measured for each subsegment. Each time the crankshaft reaches a predetermined angular position (the upper bound of each subsegment), a synchronized pressure value is computed based on the measured MAP pressures stored since the last predetermined angular position (lower bound of the current subsegment).

[0026] In the inventive process, the detection of the occurrence of backfire is made by comparing synchronized pressure values determined for a given segment (i.e. cylinder) to one or more thresholds.

[0027] Possible tests using these thresholds are presented below.

[0028] A first test (or mode) uses a first threshold TH1 . Implementing the first test involves comparing two synchronized pressure values within a same engine segment. If a difference between such two synchronized pressure values shows a pressure increase greater than a predetermined first threshold, then it is concluded that a backfire event occurred for the corresponding cylinder.

[0029] A second test (or mode) uses a second threshold TH2. Implementing the second test involves comparing two synchronized pressure values within a same engine segment. If a difference between such two synchronized pressure values shows a pressure decrease exceeding a predetermined second threshold, then it is concluded that a backfire event occurred for the corresponding cylinder. P-DELPHI-440 / WO

[0030] For tests 1 and 2, the difference may be calculated between two consecutive synchronized intake manifold pressure values. In particular, the test may be carried out by comparing values within a same segment to the first and / or second threshold, and where a backfire is observed it is for that segment / cylinder.

[0031] A third test (or mode) uses a third threshold TH3. Implementing the third test involves comparing synchronized pressure values of a same engine segment to a reference pressure, which is a representative pressure for the previous segment. Here, it is concluded to the occurrence of a backfire event for the corresponding cylinder, if a predetermined number of the synchronized pressure values exceed the reference pressure by more than the predetermined third threshold.

[0032] A fourth test (or mode) uses a fourth threshold TH4. Implementing the fourth test involves comparing synchronized pressure values of a same engine segment to a reference pressure, which is a representative pressure for the previous segment.

[0033] Here, it is concluded to the occurrence of a backfire event if a predetermined number of the synchronized pressure values show a pressure decrease lower than the fourth threshold compared to the reference pressure.

[0034] The reference pressure may be computed as an average value of the synchronized pressure values determined for the previous segment. In such case, for the fourth test / mode, where a backfire event is observed, it may rather indicate that the backfire event occurred for the previous cylinder (for which the reference pressure is calculated).

[0035] In the context of the invention, the term backfire ‘event’ is used with respect to each test / mode, to indicate a positive outcome of the test for a given cylinder, meaning that backfire did occur.

[0036] The inventive method may be operated by implementing only one of the four tests. In such case, it is concluded to a backfire “condition” if the test detects a backfire event for a given segment.

[0037] In embodiments, the method may be operated by implementing two or more of the tests, in which case a backfire condition is confirmed only if a predetermined number of the tests conclude to a backfire event. If the prescribed amount of backfire events P-DELPHI-440 / WO is observed for the respective segment, the diagnostic strategy positively concludes to backfire having occurred in the respective cylinder for a given combustion cycle.

[0038] In embodiments, the thresholds, for some or all of the four tests / modes, may be calibrated in function of average engine speed and average intake manifold pressure for the previous segment (i.e. the segment directly preceding the segment under analysis / diagnostic).

[0039] In embodiments, a cylinder specific, session backfire counter is incremented each time a backfire condition is detected for that cylinder since engine start.

[0040] In embodiments, a cylinder-specific backfire lifetime counter is incremented each time a backfire condition is detected for the respective cylinder over the engine's lifetime.

[0041] These counters, or other means, can be implemented to record the frequency of backfire conditions. Various strategies may be implemented in the engine control unit to limit the occurrence of backfire. Those strategies can be devised by the skilled person.

[0042] According to another object, the invention relates to method of operating an internal combustion engine, comprising actuating or modifying the status of one or more engine systems in response to a determined, respectively confirmed backfire condition, or after a plurality of backfire conditions have been detected for a given cylinder over a predetermined time period.

[0043] In embodiments, in case the session backfire counter of one of the cylinders exceeds a predefined threshold, the corresponding cylinder is deactivated for a predetermined number of engine cycles.

[0044] In embodiments, in case the lifetime backfire counter of one of the cylinders exceeds a predefined threshold, the corresponding cylinder is permanently deactivated.

[0045] The invention further relates to a internal combustion engine comprising a plurality of engine cylinders with respective intake ports connected to an air intake manifold; and a gaseous fuel delivery system including a pressurized fuel source supplying gaseous fuel to a fuel rail to which a plurality of fuel injectors are coupled, one per P-DELPHI-440 / WO cylinder; and wherein a control unit is configured to implement a method as described herein.

[0046] The invention still further relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out a method as herein described.

[0047] Brief Description of the Drawings

[0048] Further details and advantages of the present invention will be apparent from the following detailed description of several not limiting embodiments with reference to the attached drawings, wherein:

[0049] Fig. 1 is a diagram of internal combustion engine with port fuel injection of gaseous fuel;

[0050] Fig. 2 is a plot of intake manifold pressure vs. time, for a nominal combustion;

[0051] Figs. 3,4 and 5 are plots of intake manifold pressure vs. time illustrating the principle of the inventive method;

[0052] Figs. 6 is a flowchart implementing backfire detection modes under the present method;

[0053] Fig. 7 is a flowchart implementing a method of operating a gaseous fuel internal combustion engine that uses the present diagnostic strategies.

[0054] Description of Preferred Embodiments

[0055] Embodiments of the present invention will now be described in the context of a hydrogen internal combustion engine. The engine 10 is schematically illustrated in Fig.1 , in a simplified manner. The engine configuration is conventional and will only be briefly described. The engine 10 comprises an engine block 12 with a plurality of cylinders 14 (only one being shown in the Figure). A piston 16 is reciprocally disposed within each cylinder 14, moveable between a bottom dead center BDC and top dead center TDC, and connected to a crankshaft 18 through a rod 20. The cylinders 14 are closed by a cylinder head 22, whereby a combustion chamber 24 is defined by the piston, cylinder and cylinder head. Reference sign 26 designates an intake valve that is opened to allow fresh air inlet into the combustion chamber P-DELPHI-440 / WO

[0056] 24 (typically during an intake stroke). An exhaust valve 28 allows opening the combustion chamber 24 towards the exhaust system (not shown) to evacuate combustion gases (typically exhaust stroke).

[0057] Each intake valve 26 is connected to an air intake manifold 30 via a respective intake port 32 provided in the cylinder head 22. In this embodiment, a fuel injector 34 (per cylinder) is arranged in so-called port fuel injection (PFI) configuration; the fuel injector 34 is configured to selectively inject / discharge predetermined fuel quantities into the intake port 32 (i.e. upstream of the intake valve 26). The gaseous fuel hence mixes with air in the intake port 32 and enters the combustion chamber 24 when the intake valve 26 is opened. The method is however also applicable in the case of direct injection, i.e. where the injectors are arranged in bores in the cylinder head that open into respective combustion chambers.

[0058] Typically, the fuel injector 34 comprises a nozzle (or valve) portion that comprises a seat member with one or more injection holes; and a valve member is arranged to be moveable between a closed position, resting on the seat member to prevent fuel injection, and an open position, raised from the seat member and hence authorizing fuel flow towards the injection holes. The fuel injector typically comprises an electromechanical actuator that is configured to move the valve member. For example, the electromechanical actuator may comprise a solenoid that generates a magnetic field capable of pulling (lifting) the valve member off the seat member. For this purpose, a magnetic armature may be provided to cooperate with valve member; for example, the valve member may include a needle shaft and the armature surrounds the latter.

[0059] The fuel injectors 34 are part of a gaseous fuel delivery system, wherein the fuel injectors 34 are coupled to a fuel rail (not shown), which is fed from a pressurized fuel gas source (not shown). It may comprise one or more cylinders / tanks containing gaseous (or liquid) fuel such as hydrogen. A supply pipe connects the pressurized fuel gas source to the fuel rail, and may include serially connected components such as a shut-off valve, a pressure regulator, which may be integrated in a common housing, forming a hydrogen regulation module, HRM. The HRM may further include one or more of a pressure relief valve, a purge valve, and a fuel filter. The pressure regulator is configured to regulate the pressure in the fuel rail in a predetermined P-DELPHI-440 / WO range, e.g. from 2 to 40 bars (other pressure ranges are however possible). The lower range, e.g. 2-20 bars may be used for port fuel injection, whereas the upper range 20-40 bars may be used for direct fuel gas injection.

[0060] Back to Fig.1 , the engine typically includes a crankshaft speed sensor 40, that comprises a magnetic sensor 40.1 (e.g. hall effect sensor) in conjunction with a toothed wheel 40.2. As is known in the art, the toothed wheel 40.2 is fixed to the crankshaft, whereas the sensor 40.1 is fixed to the engine block 12 and detects changes in the magnetic field as the wheel 40.2 rotates.

[0061] The magnetic sensor 40.1 detects the presence (tooth) and absence (gap) of the metal teeth, generating a voltage pulse each time a tooth passes. The frequency of these pulses corresponds to the rotational speed of the crankshaft.

[0062] The ECU receives these pulses and processes them to determine the exact position of the crankshaft and its rotational speed.

[0063] Reference sign 42 indicates a pressure sensor, hereinafter MAP sensor, that is arranged to sense the air pressure in the intake manifold 30.

[0064] Engine operation is conventionally controlled by an engine control unit ECU (including a processor and a memory) that receives signals from various sensors and operates engine systems / components according to predetermined strategies. The ECU receives the pulses from magnetic sensor 40.1 and processes them to determine the exact position of the crankshaft and its rotational speed. The ECU further receives the signal from a pressure sensor 42 to determine values (or data) representative of the air pressure in the intake manifold.

[0065] As explained hereinabove, with hydrogen engines there is a higher risk of backfire, and it is desirable to be able to detect such defects as soon as possible.

[0066] < Invention >

[0067] The present invention proposes a method of detecting backfire in an engine, which exploits the intake manifold pressure. The present method can be programmed in the ECU to be implemented by the latter. P-DELPHI-440 / WO

[0068] In summary, the proposed method comprises a backfire diagnostic strategy that comprises the steps of: a) monitoring intake manifold pressure and engine rotation; b) determining a plurality of synchronized intake manifold pressure values over corresponding engine segments, where an engine segment represents a predefined angular range corresponding to a given cylinder; c) determining the occurrence of a backfire condition by comparing one or more of said synchronized pressure values to at least one threshold.

[0069] Fig.2 is a plot of the intake manifold pressure in function of time. MAP sensor 42 generates a sensor signal, which is indicative of the pressure PMAP in the intake manifold 30. The MAP sensor signal is sampled at a predetermined frequency, and the corresponding pressure values are stored; these measured pressure values are referred to as PMAP.

[0070] The dots in Fig.2 represent the measured pressure values PMAP. It may be observed that the dots PMAP describe a generally sinusoidal line of rather low amplitude, which is characteristic of a nominal combustion without backfire events. That is, the pressure PMAP in the intake manifold varies during engine running, but pressure variations are rather small, due to the successive intake strokes, where fresh air is drawn into the engine cylinders.

[0071] By contrast, a backfire event will cause a significant pressure variation in the intake manifold 30, which can be observed from the MAP sensor signal. The large peak in Fig.3 corresponds to a backfire event.

[0072] It may be noted that the pressure signal is sampled in an asynchronous manner, i.e. the sampling points PMAP are not dependent on the crankshaft angular position, but triggered by a timer.

[0073] In the present method, the synchronized pressure values Ps are determined based on the crankshaft signal, and can thus be assigned to predetermined engine segments, and namely to specific positions within the segment. P-DELPHI-440 / WO

[0074] Preferably, the predetermined engine segments correspond to combustion strokes (or power strokes). Hence, for an engine having N cylinders, the interval between combustion events for each cylinder is calculated as 720 / N degrees of crankshaft rotation.

[0075] Combustion segments Sc,n are represented in Figs. 2 and 3. Again, each segment Sc,n has an angular width of 7207N (i.e. 180° for a 4-cylinder engine) and corresponds to combustion in a given cylinder.

[0076] For each engine segment Sc,n, a plurality of synchronized pressure values Psare computed at predetermined crank angles 0P, the crankshaft position being known from the crankshaft sensor 40. The synchronized pressure values Ps are represented by the triangles in the Figs. The vertical arrows indicate corresponding angular positions of the crankshaft.

[0077] A synchronized pressure value Ps corresponds to the average of the MAP sensor measurement points between two consecutive, predetermined angular points 0P. So, one synchronized pressure value Ps is computed each time the crankshaft reaches a predetermined crank angle 0P.

[0078] These synchronized pressure values Ps are then used to carry out a plurality of tests, referred to as modes.

[0079] Referring to the example of Fig.2, there are here 16 samples of pressure value PMAP for each segment Sc. As indicated, the sampling is done on a time basis. When the crankshaft reaches predetermined angular positions 0Pa synchronized pressure value Ps is computed as an average of the measured (sampled) values PMAP since the previous predetermined crank angle 0P.

[0080] So, referring to the example, the predetermined angular positions 0Pare spaced by 45°. Accordingly, there are four sub-segments each with four samples PMAP. Considring segment Scjicovering the angle 180°-360°, when the crankshaft reaches 0P=225°, a first synchronized pressure value Ps is computed that represents the average of the four samples PMAP between 180 and 225°. This value Ps is represented by the triangle and positioned in the Figure in the center of the subsegment. Each time the crankshaft reaches a predetermined crank angle 0P, i.e. P-DELPHI-440 / WO every 45°, a new value of Ps is computed as the average of the pressure values PMAP measured for the segment [0p-45° ; 0P],

[0081] This can be summarized as follows: where Z is the total number of pressure samples within a subsegment [0p-45 ; 0P],

[0082] The synchronized pressure values Ps determined in accordance with the present method are assigned to a given Sc, i.e. to a given cylinder.

[0083] These synchronized pressure values Ps are then used to assess the occurrence of backfire events based on predetermined detection modes (or strategies). In practice, each mode is applied each time a new Ps value is computed.

[0084] Mode 1

[0085] This mode, which will be explained in more detail with respect to Fig.3, evaluates the pressure variation in relation to one cylinder (hence only one segment Sc,n is represented). The determination of the sampling points PMAP and of the synchronized points Ps is similar to what has been explained with regard to Fig.2.

[0086] In this first mode, a pressure variation between two consecutive (subsequent) values Ps is compared to a threshold TH1 to detect a backfire event for the observed segment Sc,n.

[0087] More specifically, Ps,i is compared to the previous value Psj-1. If the pressure difference shows an increase exceeding predetermined threshold TH1 , then it is concluded to the occurrence of backfire event in the corresponding cylinder.

[0088] This condition can be written as Ps,i > Psj-1 + THi [Test 1 ]

[0089] An exemplary flowchart for this first mode is shown in Fig. 6. Starting at step 50, suppose that a predetermined crank angle 0p,i has been reached, e.g. 0p=27O°. A table, tablel , is reset. In the method, MAP pressure and crankshaft rotation are continuously monitored. P-DELPHI-440 / WO

[0090] The MAP pressure signal is sampled (step 52) in a time base scheduler and pressure values PMAP are stored in tablel . The time base for the sampling can be adjusted e.g. between 0.1 and 2.0 ms.

[0091] An intermediate scheduler is defined to trigger the angular based pressure average measurements Ps when the crank angle reaches the predetermined angular positions 0P(O° modulo 45° in the example of the Figs). In general, the angular delta between two angular positions 0Pmay be adjusted from 5 to 45 degrees, depending on ECU capability. The values Ps are stored in table2.

[0092] So, from the angle 0pthe sampled values PMAP are stored in tablel . When the crank angle reaches 0p, j+i (test 54), the synchronized pressure value Ps is computed (step 56) as an average of the sampled values PMAP stored in table 1 (since 0Pj). Ps is stored in table2 in relation with the corresponding segment Sc,n.

[0093] The computed value Ps is then compared to threshold TH1 in test 58.

[0094] If Ps > Ps,i-i + THi, then it is concluded to a backfire event. In such case, a first backfire event flag, flag 1 , is set at step 60 and / or a backfire counter is incremented. Preferably, this test is carried out for all Ps values within the same segment Sc.

[0095] Accordingly, in the described example, at step 50 a first Ps value is already available as it has already been computed for the first sub-segment [180°; 225°]. And at step 56 a new value Ps is computed for second sub-segment [225°; 270°]. Then at step 58 these two Ps values are used in the comparison.

[0096] Threshold TH1 is represented graphically in Fig.3 It can be observed that the difference between the first two Ps values is greater than TH1. Hence in this example the test in box 58 (testl ) would be met, whereby in step 60 flagl would be raised, indicating a detected backfire event.

[0097] For this mode, it is preferably sufficient that [Test 1 ] is met for one Ps value of a segment to conclude to a backfire event.

[0098] If mode 1 is operated alone for backfire detection (i.e. without the following modes), then the outcome of a backfire event for a given segment means that a backfire condition has been detected for the corresponding cylinder. P-DELPHI-440 / WO

[0099] Mode 2

[0100] This test is designed to check whether the synchronized pressure value Ps has decreased by more than a second predetermined threshold TH2 compared to the previous pressure value Psj-i. If the difference is mathematically smaller than threshold TH2, then it is concluded to the occurrence of backfire.

[0101] This can be expressed as Ps,i < Ps -i - TH2 [test 2]

[0102] The flowchart given for mode 1 (Fig.6) is applicable. The test in box 58 being however [test 2],

[0103] In the example of Fig.3, taking for example the Ps values corresponding to the second and third sub-segments, it can be observed that the Ps value decreases by more than TH2. Hence, the test of step 58 would be met, whereby in step 60 a second flag, flag2 would be raised, indicating a detected backfire event.

[0104] For this mode, it is preferably sufficient that [Test 2] is met for one Ps value of a segment to conclude to a backfire event.

[0105] It may generally be noted that the thresholds TH1 and TH2 may be both defined by calibration based on engine speed (RPM) and average intake manifold pressure for the previous segment (Sc,n-i). The thresholds TH3 and TH4 used in the following tests may be calibrated similarly.

[0106] Mode 3

[0107] This strategy looks at pressure evolution between two consecutive cylinders, i.e. two adjacent segments Sc.

[0108] If a given number of synchronized pressure values Ps for the current cylinder have increased by more than a predetermined threshold TH3 compared to a reference pressure for the previous cylinder, then it is concluded to the occurrence of a backfire event. P-DELPHI-440 / WO

[0109] The reference pressure, noted Ps,ref, for the previous cylinder / segment is the average of the synchronized pressure values Ps measured for that previous segment (directly preceding segment).

[0110] Referring to Fig.4, Ps.ref is the average of the four Ps values determined for segment Sc,n-1 ■

[0111] For each Ps value of segment Sc,n, it is checked whether

[0112] Ps > Ps.ref + TH3 [test3]

[0113] A backfire event is found if [test3] is met for a predetermined number of Ps values within the segment Sc,n.

[0114] In the example of Fig.4, it can be observed that the difference between the first and last Ps values in segment Sc,n and Ps,ref is lower than TH3. The test [test3] would thus not be met.

[0115] However, this test is met for the second and third Ps values of Sc,n.

[0116] In this embodiment, a backfire event for segment Sc,n is found if test3 is met for a predefined number of pressure values Ps in the current segment Sc,n. This predefined number can be set between k=1 to K, where K is the total number of Ps within the segment.

[0117] Referring to the example of Fig.4, a counter C3 is incremented each time the test [test3] is met. As can be seen, the threshold TH3 is exceed by the 2ndand 3rdPs values. Supposing k=2, we have C3=2, i.e. C3>k, which means that a backfire event has occurred for the corresponding cylinder.

[0118] The status of counter C3 is indicated below the graph in Fig.4.

[0119] Mode 4

[0120] According to this strategy, if a given number of Ps for the current cylinder / segment (Sc,n) have decreased by more than a predetermined threshold TH4 compared to an average pressure Ps,ref for the previous cylinder (Sc,n-1 ), then it is concluded to the occurrence of backfire. P-DELPHI-440 / WO

[0121] This is very similar to Mode 3. For each Ps value of segment Sc,n, it is checked whether

[0122] Ps,i < Ps.ref - TH4[test4]

[0123] A backfire event is found if [test4] is met for a predetermined number of Ps values within the segment Sc,n.

[0124] Referring to the example of Fig.5, and supposing k=2 as for mode3, we have here a counter C4 that is equal to 2 at the end of the segment. Indeed, the 3rdand 4thPs values exceed TH4. It can thus be concluded that backfire has occurred (C4>k).

[0125] It may however be noted that, compared to Mode 3, the comparison to TH4 tends to indicate that pressures Ps for segment Sc,n return to standard values from an excessive reference pressure Ps,ref. Accordingly, this mode rather detects a backfire event that has occurred for the previous cylinder Sc,n-1.

[0126] Applications

[0127] In practice, depending on the engine configuration, it is possible to signal backfire on a cylinder if one of the four modes detects a backfire event.

[0128] Alternatively, a backfire condition on a respective cylinder may only be confirmed if there is convergence between two or more of the four modes. For example, the ECU may consider that a backfire situation is confirmed only if two or three of the present modes found a backfire event.

[0129] For the purpose of onboard monitoring, the ECU may comprise for each cylinder counters to count positively determined backfire conditions. For example, a counter BKFIREcyi ,tot may be configured to count all of the detected backfire conditions for a given cylinder, for the engine lifetime. BKFIREcyi, tot is thus incremented each time a backfire condition is determined for this cylinder.

[0130] A counter BKFIREcyi, Run may be configured to count the number of detected backfire conditions for a respective cylinder since engine start.

[0131] Respective thresholds may then be set to trigger actions where a BKFIREcyi, tot and / or BKFIREcyi, Run exceed their respective threshold. P-DELPHI-440 / WO

[0132] Fig.7 is an exemplary flowchart of a method of operating an internal combustion engine according to the present disclosure.

[0133] The method, starting at 100, is implemented during engine runtime, for each cylinder, and uses the diagnostic strategies discussed above in step 102. In this embodiment, step 102 implements all four modes, i.e. testl , test2, test3 and test4. If for a given segment Sc,n, two or more of these tests are positive, i.e. we have at least two tests / modes detecting a backfire event, then it is concluded to the occurrence of a backfire condition for the combustion cycle having occurred in the diagnosed cylinder. The combination of tests / modes required to conclude to a backfire condition is configurable and can be desirably set by the skilled person. For example, one, two or more of the modes / tests may need to indicate a backfire event; or for example it could be testl with any other of the tests.

[0134] Diamond 104 checks whether the output of the diagnostic step 102 is conclusive to a backfire condition. If yes, then the backfire counters for relevant cylinder(s) are incremented at 106:

[0135] BKFIREcyi.tot = BKFIREcyi.tot +1

[0136] BKFIREcyi, Run—BKFIREcyi, Run +1 (this counter being reset at each engine start)

[0137] At step 108 it is checked whether BKFIREcyi.tot has reached a predetermined threshold THB1. If yes, then the concerned cylinder is permanently deactivated at step 112. An alert (sound and / or message) may also be provided to the driver to indicate so and that he should have the combustion system checked and / or stop the engine.

[0138] At step 110 it is checked whether BKFIREcyi, Run has reached a predetermined threshold THB2. If yes, then the concerned cylinder is deactivated for a predetermined number of engine cycles, at step 114. An alert (sound and / or message) may also be provided to the driver to indicate that he should have the combustion system checked.

Claims

P-DELPHI-440 / WOClaims1. A method of detecting backfire in a gaseous fuel internal combustion engine, said engine having a plurality of cylinders with respective intake ports connected to an air intake manifold, and one fuel injector per cylinder, said method comprising: monitoring intake manifold pressure and engine rotation; determining a plurality of synchronized intake manifold pressure values over corresponding engine segments, where an engine segment represents a predefined angular range corresponding to a given cylinder; determining the occurrence of a backfire condition in a respective cylinder by comparing one or more of said synchronized pressure values to at least one threshold (TH1 , TH2, TH3, TH4).

2. The method according to claim 1 , wherein comparing comprises performing a first test, whereby it is concluded to a backfire event if a difference between two synchronized intake manifold pressure values within a same engine segment shows a pressure increase greater than a predetermined first threshold.

3. The method according to claim 1 or 2, wherein comparing comprises performing a second test, whereby it is concluded to a backfire event if a difference between two synchronized intake manifold pressure values of within a same segment shows a pressure decrease exceeding a predetermined second threshold.

4. The method according to claim 2 or 3, wherein said difference is calculated between two consecutive synchronized intake manifold pressure values.

5. The method according to any of the preceding claims, wherein comparing comprises performing a third test, whereby it is concluded to a backfire event if a predetermined number of synchronized intake manifold pressure values within a same engine segment exceed a reference pressure by more than a predetermined third threshold, said reference pressure being a representative pressure for the previous segment.P-DELPHI-440 / WO6. The method according to any of the preceding claims, wherein comparing comprises performing a fourth test, whereby it is concluded to a backfire event if a predetermined number of synchronized intake manifold pressure values of a same segment show a pressure decrease lower than a predetermined fourth threshold compared to a reference pressure being a representative pressure for the previous segment.

7. The method according to any of claims 5 or 6, wherein said reference pressure is computed as an average value of the synchronized pressure values determined for the previous segment.

8. The method according to claims 2 to 7, wherein a backfire condition is confirmed where one, two or more of said tests conclude to a backfire event.

9. The method according to any of claims 1 to 8, wherein said thresholds are calibrated in function of average engine speed and average intake manifold pressure for the previous segment.

10. The method according to any of the preceding claims, wherein said engine segment corresponds to a combustion event of the respective cylinder.11 . The method according to claim 10, wherein said engine segment has an angular width of 720° / N, where N is the number of engine cylinders.

12. The method according to any of the preceding claims, wherein monitoring said intake manifold pressure comprises sampling a sensor signal obtained from a pressure sensor arranged to sense intake manifold pressure.

13. The method according to any of the preceding claims, wherein said engine rotation is monitored by means of a crankshaft sensor.

14. The method according to any of the preceding claims, wherein the synchronized intake manifold pressure values are computed based on intake manifold pressure data measured over a predetermined crank angle interval.P-DELPHI-440 / WO15. The method according to claim 14, wherein the synchronized intake manifold pressure values are computed as an average of the intake manifold pressure data measured over a predetermined crank angle interval.

16. The method according to any of the preceding claims, wherein a cylinder specific, session backfire counter is incremented each time a backfire condition is detected for that cylinder since engine start.

17. The method according to any of the preceding claims, wherein a cylinderspecific backfire lifetime counter is incremented each time a backfire condition is detected for the respective cylinder over the engine's lifetime.

18. A method of operating an internal combustion engine, comprising actuating or modifying the status of one or more engine systems in response to a determined, respectively confirmed backfire condition, or after a plurality of backfire conditions have been detected for a given cylinder over a predetermined time period.

19. The method according to claim 18 depending on claim 16, wherein in case the session backfire counter of one of the cylinders exceeds a predefined threshold, the corresponding cylinder is deactivated for a predetermined number of engine cycles.

20. The method according to claim 18 depending on claim 16, wherein in case the lifetime backfire counter of one of the cylinders exceeds a predefined threshold, the corresponding cylinder is permanently deactivated.

21. An internal combustion engine comprising a plurality of engine cylinders with respective intake ports connected to an air intake manifold; and a gaseous fuel delivery system including a pressurized fuel source supplying gaseous fuel to a fuel rail to which a plurality of fuel injectors are coupled, one per cylinder; and wherein a control unit is configured to implement the method according to any of claims 1 to 20.P-DELPHI-440 / WO22. A computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of any of claims 1 to 20.

Images