SE543732C2

SE543732C2 - Method of performing Misfire Diagnosis, Control Arrangement, Internal Combustion Engine, and Related Devices

Info

Publication number: SE543732C2
Application number: SE1950958A
Authority: SE
Inventors: Moreno Carlos Jorques; Ola Stenlåås; Per TUNESTÅL
Original assignee: Scania Cv Ab
Priority date: 2019-08-22
Filing date: 2019-08-22
Publication date: 2021-07-06
Also published as: SE1950958A1; DE112020003386T5; WO2021034259A1

Abstract

A method (100) is disclosed of performing one or more in-cycle misfire diagnoses of the combustion of an internal combustion engine (1). The method (100) comprises obtaining (110) at least a first input value indicative of a current combustion state in a current combustion cycle, performing (120) a diagnosis of the presence of a misfire during the current combustion cycle based on the at least first input value using one or more calculation variables, evaluating (130) the diagnosis by comparing the diagnosis and the progress of the at least first input value during the current combustion cycle, and updating (140) the one or more calculation variables based on the evaluation of the diagnosis. The present disclosure further relates to a computer program, a computer-readable medium (200), a control arrangement (9), an internal combustion engine (1), and a vehicle (60).

Description

1 Method of performing l\/lisfire Diagnosis, Control Arrangement, lnternalCombustion Engine, and Related Devices TECHNICAL FIELD The present disclosure relates to a method of performing one or more in-cycle misfirediagnoses of the combustion in a cylinder of an internal combustion engine. The presentdisclosure further relates to a computer program, a computer-readable medium, a controlarrangement for an internal combustion engine, an internal combustion engine, and a vehicle comprising an internal combustion engine.

BACKGROUND Internal combustion engines, such as four-stroke internal combustion engines, comprise oneor more cylinders and a piston arranged in each cylinder. The pistons are connected to acrankshaft of the engine and are arranged to reciprocate within the cylinders upon rotation ofthe crankshaft. The engine usually further comprises one or more inlet valves and outletvalves as well as one or more fuel supply arrangements. The one or more inlet valves andoutlet valves are controlled by a respective valve control arrangement usually comprising oneor more camshafts rotatably connected to a crankshaft of the engine, via a belt, chain, gears,or similar. A four-stroke internal combustion engine completes four separate strokes whileturning a crankshaft. A stroke refers to the full travel of the piston along the cylinder, in eitherdirection. The uppermost position of the piston in the cylinder is usually referred to as the topdead centre TDC, and the lowermost position of the piston in the cylinder is usually referredto as the bottom dead centre BDC.

Multiple injections of fuel are used in some engines for reducing emissions and combustionnoise. A pilot injection is an injection of fuel performed prior to a main injection of fuel. A pilotinjection reduces the ignition delay of the main injection and reduces the heat releasemagnitude of the main combustion, such as the peak of the initial premixed combustion, andhence also the combustion noise. However, normal variations of the operating conditions,component tolerances, and aging may result in the lack of combustion occurring as a resultof the pilot injection, i.e. a pilot misfire. Moreover, different operating conditions, with specialattention to cold start, and variation of the fuel properties, may result in pilot misfires.Furthermore, pilot misfire may result also in the lack of the main combustion. The result islower indicated thermal efficiency, higher emissions, and louder combustion noise. Moreover,if the fuel is not burning, it can damage the engine and an exhaust aftertreatment system ofthe engine. 2 Furthermore, production tolerances between injectors, aging, rail pressure oscillations, andnormal variations increase the uncertainty of the injected fuel mass for a constant on-time.These variations further increase the uncertainty of pilot misfire. The effect of pilot misfire isdependent on operating conditions, and despite it has a less significant effect on the heatrelease shape at high loads, it still has a significant effect on emissions. Typically, the risk ofpilot misfire is reduced by increasing the nominal pilot mass. However, this also reduces thenominal indicated thermal efficiency.

Misfire diagnosis is required by some legislations. For example, misfire must be diagnosed inone or more cylinders when it exceeds 5%, evaluated in 1000 cycles, according to a CARBlegislation for heavy-duty engines. Misfire diagnoses are typically performed on overallcombustion of several combustion cycles. Due to above given reasons, and due to otherdifficulties, such as signal noise, it can be difficult to provide accurate diagnoses of the combustion in a cylinder of an internal combustion engine.

SUMMARYlt is an object of the present invention to overcome, or at least alleviate, at least some of theabove-mentioned problems and drawbacks.

According to a first aspect of the invention, the object is achieved by a method of performing one or more in-cycle misfire diagnoses of the combustion in a cylinder of an internal combustion engine, wherein the method comprises: - obtaining at least a first input value indicative of a current combustion state in acurrent combustion cycle in the cylinder, - performing a diagnosis of the presence of a misfire during the current combustioncycle based on the at least first input value using one or more calculation variables, - evaluating the diagnosis by comparing the diagnosis and the progress of the at leastfirst input value during the current combustion cycle, and - updating the one or more calculation variables based on the evaluation of thediagnosis.

Since the method comprises the steps of evaluating the diagnosis during the currentcombustion cycle and updating the one or more calculation variables based on theevaluation of the diagnosis, a fast, robust, and accurate diagnosis is provided of thepresence of a misfire during the current combustion cycle. 3 As a result of these features, conditions are provided for a fast, robust, and accurate controlof combustion parameters of the internal combustion engine during the current combustioncycle based on the diagnosis. Moreover, since the method comprises the step of updatingthe one or more calculation variables based on the evaluation of the diagnosis, a self-adapting method is provided minimizing calibration efforts. That is, since the one or morecalculation variables are updated based on the evaluation of the diagnosis, even more robustand accurate diagnoses can be performed in subsequent combustion cycles based on updated one or more calculation variables.

Accordingly, due to these features, the method provides conditions for improving efficiency and reducing emissions and combustion noise of internal combustion engines.

Accordingly, a method is provided overcoming, or at least alleviating, at least some of theabove-mentioned problems and drawbacks. As a result, the above-mentioned object isachieved.

I3êhe step of performing the diagnosis comprises the step of: - performing the diagnosis of the presence of a misfire during, and/or after, a pilotinjection and before the start of a time interval in which a following combustion canoccur as a result of a second pilot injection, or a main injection, of fuel into the cylinder in the current combustion cycle.

Thereby, conditions are provided for adapting the second pilot injection, and/or the maininjection, of fuel into the cylinder in the current combustion cycle based on the diagnosis. lnthis manner, a fast adaptation can be made in response to the diagnosis to thereby improve efficiency and reduce emissions and combustion noise of the internal combustion engine. 1he step of evaluating the diagnosis comprises the step of: - evaluating the diagnosis by comparing the diagnosis and the progress of the at leastfirst input value during a time interval in which a following combustion can occur as aresult of a second pilot injection, or a main injection, of fuel into the cylinder in the current combustion cycle.

Thereby, an accurate evaluation of the diagnosis can be performed and thus also anaccurate update of the one or more calculation variables based on the evaluation of thediagnosis. As a further result thereof, even more robust and accurate diagnoses can be 4 performed in subsequent combustion cycles based on updated one or more calculation variables.

Optionally, the step of updating the one or more calculation variables comprises the steps of: - determining the accuracy of the diagnosis by comparing the diagnosis and theprogress of the at least first input value, and - updating the one or more calculation variables if the accuracy of the diagnosis isbelow a threshold accuracy.

Thereby, the one or more calculation variables is/are updated in a robust and accuratemanner. As a result thereof, even more robust and accurate diagnoses can be performed in subsequent combustion cycles based on updated one or more calculation variables.

Optionally, the one or more calculation variables is/are one or more thresholds, and wherein the step of performing the diagnosis comprises the step of: - performing the diagnosis of the presence of a misfire by comparing the at least firstinput value and the one or more thresholds.

Thereby, a simple, robust, and accurate method is provided requiring little computationalresources for performing the diagnosis of the presence of a misfire during the currentcombustion cycle.

Optionally, the step of updating the one or more calculation variables comprises the step of:- updating the one or more calculation variables based on the difference between theat least first input value and the one or more thresholds.

Optionally, the step of updating the one or more calculation variables comprises the step of: - updating the one or more calculation variables with a magnitude being substantiallyinversely proportional to the difference between the at least first input value and theone or more thresholds.

Larger differences between the at least first input value and the one or more thresholds maybe the result of temporary deviations in the obtained at least first input value. However, since the one or more calculation variables is/are updated with a magnitude being substantiallyinversely proportional to the difference between the at least first input value and the one ormore thresholds, larger differences will have a lower impact on the updated calculationvariables than smaller differences. Thus, due to these features, the one or more calculationvariables is/are updated in a further robust and accurate manner.

Optionally, the one or more calculation variables is/are one or more misfire probabilities.Thereby, an even more robust and accurate diagnosis can be performed. l\/loreover,conditions are provided for updating the one or more calculation variables also in caseswhere the evaluation of the diagnosis shows that the diagnosis was accurate. ln this manner,even more robust and accurate diagnoses can be performed based on updated one or more calculation variables.

Optionally, the method comprises the step of:- obtaining at least one misfire probability of the one or more misfire probabilities bymodelling the probability of a misfire using the at least first input value.

Thereby, an even more robust and accurate diagnosis can be performed. l\/loreover,conditions are provided for comparing different probabilities and selecting the most accurateprobability. Furthermore, conditions are provided for updating the diagnosis with improvedpartial models whose probabilities easily can be compared with other partial models.

Optionally, the method comprises the step of:- obtaining at least one misfire probability of the one or more misfire probabilities bycomputing the probability of a misfire with a stochastic model using the at least first input value.

Thereby, an even more robust and accurate diagnosis can be performed.

Optionally, the method comprises the step of:- adapting combustion parameters of the internal combustion engine during the currentcombustion cycle based on the diagnosis.

Thereby, a fast, robust, and accurate adaptation of combustion parameters is performedduring the current combustion cycle based on the diagnosis. ln this manner, the occurrencesof misfire can be reduced. Accordingly, due to these features, a method is provided capable 6 of improving efficiency and reducing emissions and combustion noise of internal combustion engines.

Optionally, the step of performing the diagnosis comprises the steps of: - performing a first separate diagnosis of the presence of a misfire in the currentcombustion cycle based on the at least first input value using one or more calculationvariables, - obtaining at least a second input value indicative of the current combustion state inthe current combustion cycle, - performing a second separate diagnosis of the presence of a misfire in the currentcombustion cycle based on the at least second input value using one or morecalculation variables, and - combining the first and second separate diagnoses into the diagnosis of the presence of a misfire in the current combustion cycle.

Thereby, an even more robust and accurate diagnosis can be performed. This because atleast two different input values are used to perform two separate diagnoses which arecombined into the diagnosis of the presence of a misfire in the current combustion cycle. lnthis manner, an even more accurate control of combustion parameters of the internalcombustion engine can be performed during the current combustion cycle based on thediagnosis.

Optionally, the method comprises the steps of: - evaluating the first separate diagnosis by comparing the first separate diagnosis andthe progress of the at least first input value during the current combustion cycle, - providing a first weighting factor based on the evaluation of the first separatediagnosis, - evaluating the second separate diagnosis by comparing the second separatediagnosis and the progress of the at least second input value during the currentcombustion cycle, and - providing a second weighting factor based on the evaluation of the second separatediagnosis, and wherein the method comprises the step of, during the current combustion cycle and/or during a subsequent combustion cycle: - combining the first and second separate diagnoses into the diagnosis of the presenceof a misfire using the first and second weighting factors. 7 Thereby, an even more robust and accurate diagnosis can be performed. This because atleast two different input values are used to perform two separate diagnoses which areevaluated and combined into the diagnosis of the presence of a misfire in the currentcombustion cycle using the first and second weighting factors. ln this manner, an even moreaccurate control of combustion parameters of the internal combustion engine can beperformed during the current combustion cycle and/or during a subsequent combustion cyclebased on the diagnosis.

Optionally, the at least first input value is indicative of one of a position of heat release, a current heat release, a current accumulated heat release, and a current cylinder pressure.

Thereby, a robust and accurate diagnosis can be performed of the presence of a misfireduring the current combustion cycle.

Optionally, the at least second input value is different from the at least first input value and isindicative of one of a position of heat release, a current heat release, a current accumulatedheat release, and a current cylinder pressure.

Thereby, an even more robust and accurate diagnosis can be performed of the presence of a misfire during the current combustion cycle.

Optionally, the method comprises: - inputting one or more input values indicative of a current operational state of thecombustion engine, and - adapting the one or more calculation variables based on the current operational stateof the combustion engine.

According to a second aspect of the invention, the object is achieved by a computer programcomprising instructions which, when the program is executed by a computer, cause thecomputer to carry out the method according to some embodiments of the present disclosure.Since the computer program comprises instructions which, when the program is executed bya computer, cause the computer to carry out the method according to some embodiments, acomputer program is provided which provides conditions for overcoming, or at least alleviating, at least some of the above-mentioned drawbacks. As a result, the above-mentioned object is achieved.

According to a third aspect of the invention, the object is achieved by a computer-readablemedium comprising instructions which, when executed by a computer, cause the computer tocarry out the method according to some embodiments of the present disclosure. Since thecomputer-readable medium comprises instructions which, when executed by a computer,cause the computer to carry out the method according to some embodiments, a computer-readable medium is provided which provides conditions for overcoming, or at leastalleviating, at least some of the above-mentioned drawbacks. As a result, the above-mentioned object is achieved.

According to a fourth aspect of the invention, the object is achieved by a control arrangement for an internal combustion engine, wherein the control arrangement is configured to perform one or more in-cycle misfire diagnoses of the combustion in a cylinder of the internal combustion engine, wherein the control arrangement is configured to: - obtain at least a first input value indicative of a current combustion state in a currentcombustion cycle in the cylinder, - perform a diagnosis of the presence of a misfire during the current combustion cyclebased on the at least first input value using one or more calculation variables, - evaluate the diagnosis by comparing the diagnosis and the progress of the at leastfirst input value during the current combustion cycle, and - update the one or more calculation variables based on the evaluation of thediagnosis.

Performing the diagnosis gginnrrises: - performing the diagnosis of the oreseitoe of a inisfire durino. ahdfor after a niiotiniection and before the start of a time iritervai in trrritioh a toiioiriiirtg oombtistiori oartoccur as a resuit of a second oiiot inieotiort. or a main inieotiori, oi ttiei into the gviiitder in the current combustion ovgie.

Evaiuatino the diagnosis oorrtortses: - exfairratino tite diagnosis by oontoaring tne diagnosis and the orogress of the at ieast first ingut xfaiue during a time intervai in tuitigh a toiigtrrind combustion can occur as a resuit ot a second piiot inieotion. or a inain iniectigh. ot tue! into the gviiitoer in the current oomotistiort cyeie. 9 Since the control arrangement is configured to evaluate the diagnosis during the currentcombustion cycle and update the one or more calculation variables based on the evaluationof the diagnosis, a control arrangement is provided capable of performing a fast, robust, andaccurate diagnosis of the presence of a misfire during the current combustion cycle.

As a result of these features, conditions are provided for a fast, robust, and accurate controlof combustion parameters of the internal combustion engine during the current combustion cycle based on the diagnosis.

Moreover, since the control arrangement is configured to update the one or more calculationvariables based on the evaluation of the diagnosis, a self-adapting control arrangement isprovided minimizing calibration efforts. That is, since the one or more calculation variablesare updated based on the evaluation of the diagnosis, a control arrangement is providedcapable of performing even more robust and accurate diagnoses in subsequent combustioncycles based on updated one or more calculation variables.

Accordingly, due to these features, the control arrangement provides conditions for improving efficiency and reducing emissions and combustion noise of internal combustion engines.

Accordingly, a control arrangement is provided overcoming, or at least alleviating, at leastsome of the above-mentioned problems and drawbacks. As a result, the above-mentionedobject is achieved.

According to a fifth aspect of the invention, the object is achieved by an internal combustionengine comprising a control arrangement according to some embodiments of the presentdisclosure.

Since the internal combustion engine comprises a control arrangement according to someembodiments, an internal combustion engine is provided in which fast, robust, and accuratediagnoses can be made of the presence of misfires. As a further result thereof, conditionsare provided for improving efficiency and reducing emissions and combustion noise of the internal combustion engine.

Accordingly, an internal combustion engine is provided overcoming, or at least alleviating, atleast some of the above-mentioned problems and drawbacks. As a result, the above-mentioned object is achieved.

According to a sixth aspect of the invention, the object is achieved by a vehicle comprising aninternal combustion engine according to some embodiments of the present disclosure.

Thereby, a vehicle is provided comprising an internal combustion engine in which fast,robust, and accurate diagnoses can be made of the presence of misfires. As a further resultthereof, conditions are provided for improving efficiency and reducing emissions and noise ofthe vehicle.

Accordingly, a vehicle is provided overcoming, or at least alleviating, at least some of theabove-mentioned problems and drawbacks. As a result, the above-mentioned object isachieved.

Further features of, and advantages with, the present invention will become apparent whenstudying the appended claims and the following detailed description.

BRIEF DESCRIPTION OF THE DRAWINGS Various aspects of the invention, including its particular features and advantages, will bereadily understood from the example embodiments discussed in the following detaileddescription and the accompanying drawings, in which: Fig. 1 schematically illustrates a cross sectional view of an internal combustion engine,according to some embodiments, Fig. 2 illustrates a first graph showing heat release during a first example combustion cycle inthe cylinder of the engine illustrated in Fig. 1, Fig. 3 illustrates a second graph showing heat release during a second example combustioncycle in the cylinder of the engine illustrated in Fig. 1, Fig. 4 illustrates a vehicle, according to some embodiments, Fig. 5 illustrates a method of performing one or more in-cycle misfire diagnoses of thecombustion in a cylinder of an internal combustion engine, and Fig. 6 illustrates computer-readable medium, according to some embodiments.

DETAILED DESCRIPTION Aspects of the present invention will now be described more fully. Like numbers refer to likeelements throughout. Well-known functions or constructions will not necessarily be describedin detail for brevity and/or clarity. 11 Fig. 1 schematically illustrates a cross sectional view of an internal combustion engine 1,according to some embodiments. For the reason of brevity and clarity, the internalcombustion engine 1 is in some places herein referred to as “the combustion engine”, orsimply “the engine 1”. The engine 1 comprises at least one cylinder 3 and a piston 12arranged in each cylinder 3. The piston 12 is connected to a crankshaft 16 via a connectingrod 13. The piston 12 moves forwards and backwards in the cylinder 3 between a top deadcentre and a bottom dead centre upon rotation of the crankshaft 16. The engine 1 comprisesan inlet system 14, which in the illustrated example engine is illustrated as an inlet duct. Theinlet system 14 may further comprise an air filter, and according to some embodiments athrottle, a fuel injector, an air flow sensor, etc. ln the illustrated embodiments, the inletsystem 14 is fluidically connected to a compressor of a charging device 36, as will be furtherexplained below.

According to the illustrated embodiments, the engine 1 is a four-stroke internal combustionengine 1 which comprises at least one inlet valve 18 arranged in each cylinder 3, which atleast one inlet valve 18 is connected with the inlet system 14. The engine 1 further comprisesan inlet valve control arrangement 22 configured to control each inlet valve 18 on the basis ofa rotational position of the crankshaft 16. The engine 1 further comprises at least oneexhaust valve 24 arranged in each cylinder 3, which at least one exhaust valve 24 isconnected with an exhaust outlet 26 of the engine 1. The engine 1 further comprises anexhaust valve control arrangement 28 configured to control each exhaust valve 24 on thebasis of the rotational position of the crankshaft 16. ln Fig. 1, the at least one inlet valve 18and the at least one exhaust valve 24 are illustrated in a respective closed position. ln theclosed position, each valve 18, 24 abuts against a respective valve seat to close fluidconnection between the cylinder 3 and the respective inlet system 14 and the exhaust outlet26.

The inlet valve control arrangement 22 is arranged to control the at least one inlet valve 18between the closed position and an open position by displacing the at least one inlet valve 18in a direction into the cylinder 3. A fluid connection is thereby opened between the inletsystem 14 and the cylinder 3. Likewise, the exhaust valve control arrangement 28 isarranged to control the at least one exhaust valve 24 between the closed position and anopen position by displacing the at least one exhaust valve 24 in a direction into the cylinder3. Thereby, a fluid connection is opened between the cylinder 3 and the exhaust outlet 26.Upon displacement of a valve 18, 24 from the closed position to the open position, the valve18, 24 is lifted from its valve seat. The engine 1 further comprises a fuel injector 31 arrangedto directly inject fuel into the cylinder 3. The engine 1 in the illustrated embodiments is a 12 compression ignition engine, such as a diesel engine. According to further embodiments, theengine may be an Otto engine with a spark-ignition device, wherein the Otto engine may be designed to run on gas, petrol, alcohol or similar volatile fuels or combinations thereof. Suchfuel may be directly injected into the cylinder 3 using a fuel injector.

The exhaust valve control arrangement 28 and the inlet valve control arrangement 22 mayeach comprise one or more camshafts rotatably connected to the crankshaft 16, wherein thecamshafts comprises cam lobes arranged to displace valves 18, 24 to an open position bypressing on valve stems of the valves 18, 24 upon rotation of the camshaft. The exhaustvalve control arrangement 28 and/or the inlet valve control arrangement 22 may according tofurther embodiments comprise electric, pneumatic, or hydraulic actuators arranged to controlvalves on the basis of the rotational position of the crankshaft 16. The rotational position ofthe crankshaft 16 may be obtained using a crank angle sensor 29.

The exhaust valve control arrangement 28 comprises an exhaust valve phase-shifting device30 configured to phase-shift control of the at least one exhaust valve 24 in relation to thecrankshaft 16. Further, the inlet valve control arrangement 22 comprises an inlet valvephase-shifting device 32 configured to phase-shift control of the at least one inlet valve 18 inrelation to the crankshaft 16. The exhaust valve phase-shifting device 80 and the inlet valvephase-shifting device 82 may each comprise a hydraulic arrangement, for example usingengine oil as hydraulic fluid, to phase-shift control of the valves 18, 24 in relation to thecrankshaft 16. Such hydraulic arrangement may form part of a belt pulley (not illustrated)arranged to transfer rotation from the crankshaft 16 to a camshaft of the exhaust valvecontrol arrangement 28 and/or the inlet valve control arrangement 22, wherein the hydraulicarrangement is arranged to regulate an angular relationship between a first portion of the beltpulley, being connected to the crankshaft 16, and a second portion of the belt pulley, beingconnected to the camshaft, in order to phase-shift control of the at least one inlet valve 18and/or the at least one exhaust valve 24. ln embodiments wherein the exhaust valve controlarrangement 28 and/or the inlet valve control arrangement 22 comprises electric, pneumaticor hydraulic actuators, the phase-shift of control of the at least one inlet valve 18 and/or theat least one exhaust valve 24 may be performed in other manners, for example by anelectronic phase-shift of control.

According to the illustrated embodiments, the engine 1 comprises an exhaust after treatmentsystem 39. The exhaust after treatment system 39 may comprise one or more of a catalyticconverter, a particulate filter, a Selective catalytic reduction (SCR) arrangement, a DieselOxidation Catalyst (DOC), a Lean NOx Trap (LNT) and a Three-Way Catalyst (TWC). 13 Moreover, according to the illustrated embodiments, the engine 1 comprises a chargingdevice 36 arranged to compress air to the inlet system 14. The charging device 36 illustratedis a turbo-charger comprising a turbine arranged to be driven by gases from the exhaustoutlet 26. The turbine is arranged at a shaft connected to a compressor wheel which isarranged to compress air to the inlet system 14. The engine 1 may comprise another type ofcharging device, such as a compressor arranged to be driven by the crankshaft 16 of the engine 1.

The engine 1 comprises a sensor 5 configured to obtain at least a first input value indicativeof a current combustion state in a current combustion cycle in the cylinder 3. According to theillustrated embodiments, the sensor 5 is a cylinder pressure sensor configured to measurethe pressure in the cylinder 3. By measuring the pressure in the cylinder 3, the heat releasein the cylinder 3 can be sensed in a direct manner. According to further embodiments, theengine 1 may comprise another type of sensor configured to obtain at least a first input valueindicative of a current combustion state in a current combustion cycle in the cylinder 3, suchas an engine speed sensor, an crank angle position sensor, an crank shaft torque sensor, aknock sensor, an engine block vibration sensor, an engine top vibration sensor, an engineblock strain sensor, an engine top strain sensor, an ion-current sensor, a virtual sensor forthe cylinder pressure, a virtual sensor for the heat release, or the like. According to someembodiments, the at least first input value may be obtained using sensor fusion techniqueswhere measurements of different sensors are fused into at least first input value indicative ofa current combustion state in a current combustion cycle in the cylinder 3 of the engine 1.Such different sensors may comprise one or more of the above mentioned types of sensors.

The engine 1 comprises a control arrangement 9. As is further explained herein, the controlarrangement 9 is configured to perform one or more in-cycle misfire diagnoses of thecombustion in the cylinder 3 of the engine 1 based on the at least first input value.

Fig. 2 illustrates a first graph showing heat release Hr during a first example combustioncycle in the cylinder 3 of the engine 1 illustrated in Fig. 1. Below, simultaneous reference ismade to Fig. 1 and Fig. 2. ln Fig. 2, the heat release Hr is shown in a range from -40 crankangle degrees CAD to 30 crank angle degrees CAD. Zero crank angle degrees CAD in Fig. 2corresponds to the top dead centre of the piston 12 of the engine 1 illustrated in Fig. 1.Negative crank angle degrees CAD are crank angle degrees before the top dead centre andpositive crank angle degrees CAD are crank angle degrees after the top dead centre. Theheat release Hr in Fig. 2 is shown in Joules per crank angle degree CAD. l\/loreover, the 14 graph of Fig. 2 illustrates injection currents of a pilot injection pi1, a second pilot injection pi2,and a main injection mi of fuel into the cylinder 3 in the current combustion cycle.

As mentioned, according to the illustrated embodiments, the control arrangement 9 isconfigured to obtain the at least a first input value indicative of a current combustion state ina current combustion cycle in the cylinder 3 from the cylinder pressure sensor 5. Thus,according to the illustrated embodiments, the at least first input value is indicative of a currentheat release Hr in a current combustion cycle in the cylinder 3. Therefore, in some placesherein, the reference sign “Hr” is used for the at least first input value. According to furtherembodiments, the at least first input value may be indicative of one of a position of heatrelease Hr, a current accumulated heat release Hr, and a current cylinder pressure, or thelike, as is further explained herein. The control arrangement 9 is further configured to performa diagnosis of the presence of a misfire in the cylinder 3 during the current combustion cyclebased on the at least first input value Hr using one or more calculation variables. The one ormore calculation variables may comprise one or more thresholds and/or one or more misfireprobabilities, as is further explained herein. Moreover, the control arrangement 9 isconfigured to evaluate the diagnosis by comparing the diagnosis and the progress of the atleast first input value Hr during the current combustion cycle and is configured to update theone or more calculation variables based on the evaluation of the diagnosis. ln this manner, a control arrangement 9 is provided capable of performing a fast, robust, andaccurate diagnosis of the presence of a misfire in the cylinder 3 during the currentcombustion cycle. Moreover, since the control arrangement 9 is configured to update the oneor more calculation variables based on the evaluation of the diagnosis, a self-adaptingcontrol arrangement 9 is provided minimizing calibration efforts. That is, since the one ormore calculation variables are updated based on the evaluation of the diagnosis, a controlarrangement 9 is provided capable of performing even more robust and accurate diagnosesin subsequent combustion cycles based on updated one or more calculation variables, as isfurther explained herein.

According to the illustrated embodiments, the control arrangement 9 is configured to performthe diagnosis of the presence of a misfire during, and/or after, a pilot injection pi1 and beforethe start of a time interval Ti in which a following combustion can occur as a result of asecond pilot injection pi2, or a main injection mi, of fuel into the cylinder 3 in the currentcombustion cycle. ln this manner, the diagnosis can be performed while there is stillcontrollability of the second pilot injection pi2 and/or the main injection mi of fuel into thecylinder 3. Accordingly, due to these features, conditions are provided for adapting the second pilot injection pi2, and/or the main injection mi, of fuel into the cylinder 3 in the currentcombustion cycle based on the diagnosis, as is further explained herein.

According to the illustrated embodiments, the control arrangement 9 is configured to evaluatethe diagnosis by comparing the diagnosis and the progress of the at least first input value Hrduring the time interval Ti in which a following combustion can occur as a result of a secondpilot injection pi2, or a main injection mi, of fuel into the cylinder 3 in the current combustioncycle. The time interval Ti, as referred to herein, may also encompass a crank angle intervalin which a following combustion can occur as a result of a second pilot injection pi2, or amain injection mi, of fuel into the cylinder 3 in the current combustion cycle. As commonlyknown in the technical field, crank angle degrees CAD can be converted into time using therotational speed of the engine 1 as input. As an example, 10 crank angle degrees CADcorresponds to approximately 1.39 milliseconds when the rotational speed of the engine 1 is 1200 revolutions per minute. ln the first example combustion cycle illustrated in Fig. 2, there is a misfire of the pilotinjection pi1 as well as a misfire of the second pilot injection pi2. Thereby, the resultingcombustion pC, C1 is a combustion of the fuel injected in the pilot injection pi1, the secondpilot injection pi2 and the main injection mi. As a result thereof, as can be seen in Fig. 2, arelative great premixed combustion pC is obtained at approximately 7 crank angle degreesafter top dead centre as a result of the fuel injected in the pilot injection pi1, the second pilotinjection pi2 and the main injection mi. ln Fig. 2, the heat release Hr peaks at approximately380 Joules per crank angle degrees CAD at approximately 7 crank angle degrees after topdead centre. The relatively high maximum heat release Hr and the relatively high gradient ofthe heat release Hr can result in a lower indicated thermal efficiency, increased combustionnoise and pollutant emissions, especially NOx pollutant emissions.

Fig. 3 illustrates a second graph showing heat release Hr during a second examplecombustion cycle in the cylinder 3 of the engine 1 illustrated in Fig. 1. Below, simultaneousreference is made to Fig. 1 and Fig. 3. ln Fig. 3, the heat release Hr is shown in a range from-40 crank angle degrees CAD to 30 crank angle degrees CAD. Zero crank angle degreesCAD in Fig. 3 corresponds to the top dead centre of the piston 12 of the engine 1 illustratedin Fig. 1. Negative crank angle degrees CAD are crank angle degrees before the top deadcentre and positive crank angle degrees CAD are crank angle degrees after the top deadcentre. The heat release Hr in Fig. 3 is shown in Joules per crank angle degree CAD.Moreover, the graph of Fig. 3 illustrates injection currents of a pilot injection pi1, a second 16 pilot injection pi2, and a main injection mi of fuel into the cylinder 3 in the current combustioncycle. ln the second example combustion cycle illustrated in Fig. 3, there is a combustion G1occurring as a result of the pilot injection pi1. The pilot injection pi1 is performed atapproximately -40 crank angle degrees GAD and the combustion G1 is occurring atapproximately -28 crank angle degrees GAD. Accordingly, in the second examplecombustion cycle illustrated in Fig. 3, the ignition delay of the pilot injection pi1 isapproximately 12 crank angle degrees GAD. Moreover, in the second example combustioncycle illustrated in Fig. 3, there is a combustion G2 occurring as a result of the second pilotinjection pi2. The second pilot injection pi2 is performed at approximately -14 crank angledegrees GAD and the combustion G2 is occurring at approximately -6 crank angle degreesGAD. Accordingly, in the second example combustion cycle illustrated in Fig. 3, the ignitiondelay of the second pilot injection pi2 is approximately 8 crank angle degrees GAD.

Due to the combustion G1 occurring as a result of the pilot injection pi1 and the combustionG2 occurring as a result of the second pilot injection pi2, the main ignition delay is reducedwhich enlarges the duration of the diffusive combustion G3. ln this manner, a relativelysmaller maximum heat release Hr is obtained. As can be seen in Fig. 3, the heat release Hrpeaks at approximately 260 Joules per crank angle degrees GAD at approximately 8 crankangle degrees after top dead centre. The relatively smaller maximum heat release Hr and therelatively smaller gradient of the heat release Hr can result in a higher indicated thermalefficiency, reduced combustion noise and pollutant emissions, especially NOx pollutant emissions. ln the first example combustion cycle of Fig. 2, if the control arrangement 9 performs adiagnosis showing presence of a misfire based on the at least first input value Hr using theone or more calculation variables, the control arrangement 9 may adapt combustionparameters of the internal combustion engine 1 during the current combustion cycle basedon the diagnosis. Such adaptation of combustion parameters may include adaptation of fuelinjection of the fuel injector 31, such as a cancellation of further injections Pi2, mi, anadvance or postponing of further injections Pi2, mi, a change of amount of injected fuel infurther injections Pi2, mi, phase-shifting control of inlet valves 18 and/or exhaust valves 24,or the like.

As an alternative, or in addition, if the control arrangement 9 performs a diagnosis showingpresence of a misfire based on the at least first input value Hr using the one or more 17 calculation variables, the control arrangement 9 may adapt combustion parameters of theinternal combustion engine 1 having effect in subsequent combustion cycles of the engine 1.Such adaptations of combustion parameters may include phase-shifting control of inletvalves 18 and/or exhaust valves 24, change in charge pressure of a charging device 36 ofthe engine 1, change in rail pressure, or the like.

According to the embodiments illustrated in Fig. 1, the control arrangement 9 is connected tothe charging device 36. The control arrangement 9 may be configured to change the chargepressure of the charging device 36 of by regulating a waste gate valve of the charging device36. According to further embodiments, the charging device 36 is a Variable-geometryturbocharger (VGT). ln such embodiments, the control arrangement 9 may be configured toregulate the charge air pressure of the charging device 36 by regulating geometry of an inletportion of the turbine of the turbocharger, for example by regulating angular positions ofvanes arranged at the inlet portion of the turbine.

According to the embodiments illustrated in Fig. 1, the control arrangement 9 is connected tothe exhaust valve phase-shifting device 30 and the inlet valve phase-shifting device 32. Thecontrol arrangement 9 may be configured to phase-shifting control of inlet valves 18 and/orexhaust valves 24 by controlling the exhaust valve phase-shifting device 30 and/or the inletvalve phase-shifting device 32. ln the first example combustion cycle of Fig. 2, if the control arrangement 9 performs adiagnosis showing presence of a misfire based on the at least first input value Hr using theone or more calculation variables, the control arrangement 9 may evaluate the diagnosis bycomparing the diagnosis and the progress of the at least first input value Hr during thecurrent combustion cycle. Furthermore, the control arrangement 9 may update the one ormore calculation variables based on the evaluation of the diagnosis.

The control arrangement 9 may determine the accuracy of the diagnosis by comparing thediagnosis and the progress of the at least first input value Hr and update the one or morecalculation variables if the accuracy of the diagnosis is below a threshold accuracy. As anexample, if the control arrangement 9 performs a diagnosis showing presence of a misfirebased on the at least first input value Hr using the one or more calculation variables, and theprogress of the at least first input value Hr during the remaining part of the currentcombustion cycle indicates that the diagnosis was incorrect, the control arrangement 9 mayupdate the one or more calculation variables. The diagnosis can be determined to beincorrect, i.e. to have a low accuracy, if progress of the at least first input value Hr during the 18 remaining part of the current combustion cycle indicates a combustion occurring as a resultof a pilot injection pi1, pi2, and/or a less intense subsequent heat release Hr, as illustrated inFig. 3. Contrariwise, the diagnosis can be determined to be correct, i.e. to have a highaccuracy, if progress of the at least first input value Hr during the remaining part of thecurrent combustion cycle indicates a lack of combustion, or a relatively small combustion,occurring as a result of a pilot injection pi1, pi2, and/or a more intense subsequent heatrelease Hr, as illustrated in Fig. 2. Moreover, by comparing how similar the measured mainpremixed heat release magnitude is to each of the cases, the cycle can be classified as amisfire or combustion with higher accuracy. Furthermore, the control arrangement 9 maydetermine the accuracy of the diagnosis by analysing the timing of the maximum heatrelease Hr, and/or the difference of the ignition delay of the main injection mi.

According to some embodiments, the one or more calculation variables is/are one or morethresholds. According to such embodiments, the control arrangement 9 may perform thediagnosis of the presence of a misfire by comparing the at least first input value Hr and theone or more thresholds. As an example, in the examples illustrated in Fig. 2 and Fig. 3, thecalculation variables may comprise a threshold of 50 Joules per Joules per crank angledegrees CAD. lf the at least first input value Hr indicates a heat release during, and/or after,the pilot injection pi1 below a threshold value, the control arrangement 9 can diagnose apresence of a misfire during the current combustion cycle. Contrariwise, if the at least firstinput value Hr indicates a heat release during, and/or after, the pilot injection pi1 above thethreshold value, the control arrangement 9 can diagnose an absence of a misfire, i.e. acombustion, during the current combustion cycle.

According to some embodiments, the control arrangement 9 may be configured to update theone or more calculation variables, i.e. the one or more thresholds according to someembodiments, based on the difference between the at least first input value Hr and the oneor more thresholds. According to these embodiments, the control arrangement 9 may beconfigured to update the one or more calculation variables with a magnitude beingsubstantially inversely proportional to the difference between the at least first input value Hrand the one or more thresholds. Thereby, the one or more calculation variables is/areupdated in a further robust and accurate manner.

The one or more calculation variables can be updated every cycle a diagnosis fails, i.e. whenthe diagnosis of a cycle is determined to be incorrect, proportionally to an adaptation gain K.For a faster update, a recursive equation may use additional information provided by thedistance to the threshold. ln this way, the one or more calculation variables, i.e. the one or 19 more thresholds according to some embodiments, is/are not updated when the diagnosis isdetermined correct, but it is adjusted inversely proportional to the threshold distance whenthe in-cycle diagnosis is determined incorrect. A small constant may be added to avoid division by zero in the recursive equation. lt is to be noted that the first and second example combustion cycles i||ustrated in Fig. 2 andFig. 3 only constitute examples. For example, the engine 1 may operate using more than twopilot injections pi1, pi2, or less than two pilot injections pi1, pi2. Moreover, as explainedabove, if a misfire is diagnosed, as is the case in Fig. 2, the control arrangement 9 mayperform an action in the current combustion cycle, i.e. in the same combustion cycle,involving a cancellation of further injections pi2, mi of fuel into the cylinder 3 in the currentcombustion cycle.

As is commonly known in the technical field, combustion may occur in the cylinder 3 of anengine 1 during the compression stroke before the piston 12 reaches the top dead centreTDC as well as during the expansion stroke, i.e. after the piston 12 has passed the top deadcentre TDC. The expansion stroke is also known as the power stroke or the combustionstroke. ln the second example combustion cycle i||ustrated in Fig. 3, the combustion C1 isoccurring at approximately -28 crank angle degrees CAD and the combustion C2 is occurringat approximately -6 crank angle degrees. That is, each of these combustions C1, C2 isoccurring during the compression stroke of the piston 12 before the piston 12 reaches the topdead centre TDC. Apparently, the wording “combustion cycle”, as used herein, may compriseportions of the compression stroke and the expansion stroke of the engine 1. Thecombustion cycle, as referred to herein, may be defined as a cycle of the engine 1 in whichcombustion of injected fuel is expected to occur during the compression stroke and/or theexpansion stroke of the engine 1. The combustion cycle, as referred to herein, may rangefrom the first injection of fuel to the end of the expansion stroke of the engine 1, wherein thefirst injection of fuel is performed during the compression stroke or the expansion stroke. lnthe first and second example combustion cycles i||ustrated in Fig. 2 and Fig. 3, the firstinjection of fuel is the first pilot injection pi1, which is injected during the compression stroke,as explained above. The wording “current combustion cycle”, as used herein, refers to acurrent ongoing combustion cycle as defined above.

According to some embodiments of the present disclosure, the one or more calculationvariables is/are one or more misfire probabilities. According to some embodiments, thecontrol arrangement 9 may be configured to obtain at least one misfire probability of the oneor more misfire probabilities by modelling the probability of a misfire using the at least first input value Hr. Moreover, the control arrangement 9 may be configured to obtain at least onemisfire probability of the one or more misfire probabilities by computing the probability of amisfire with a stochastic model using the at least first input value Hr.

Furthermore, as is further explained herein, the one or more calculation variables maycomprise one or more fire probabilities. The control arrangement 9 may be configured toobtain at least one fire probability of the one or more fire probabilities by modelling theprobability of a fire, i.e. combustion, using the at least first input value Hr. l\/loreover, thecontrol arrangement 9 may be configured to obtain at least one fire probability of the one ormore fire probabilities by computing the probability of a fire with a stochastic model using theat least first input value Hr.

A Bayesian decision approach may be used to minimize the average detection errorprobability. This is equivalent to maximizing the detection performance, measured as thefraction of correctly detected pilot combustion C1 or misfire. The decision rule to optimize theaverage detection error, where x is the at least first input value, is: decide misfire: P(misfire|x) > P(fire|x) ln the following, the measured variables as referred to may comprise at least a first inputvalue and/or at least a second input value each indicative of the current combustion state inthe current combustion cycle in the cylinder 3 of the engine 1. Two approaches to obtain theprobabilities can be used. The first one is the direct modelling of the misfire probability for agiven measured value of each variable. ln the second approach, the value of each variable ispredicted together with its uncertainty. By applying Bayes rule, the posterior probability ofpilot combustion C1 and misfire can be computed. The first approach provides a systematicmethod for having a variable threshold on the measured variables with the operatingconditions. The second approach simplifies how the probability distributions are updated bycycle-to-cycle feedback and how the different calculation variables information is combined.

The misfire probability can be modelled as normally distributed, where the mean andvariance are functions of the at least one calculation variable, e.g. the heat releasemagnitude, the accumulated heat release, and/or the pressure rise over the motoring DFGSSUFG tfaCe.

Predictive stochastic models can be used in which the pilot pi1 heat release Hr magnitude,its crank angle CAD position, the pilot pi1 accumulated heat release Hr and the pressure rise 21 over motoring can be modelled. Their value and uncertainty can be predicted under the twohypotheses, pilot combustion C1, and pilot misfire. The posterior probability can then becomputed by the Bayes rule. The misfire probability is the prior probability of obtaining pilotmisfire when no more information but the injected pilot pi1 mass and the start of injection areknown. The prior probability of obtaining a measured value of x may be computed via the lawof total probability.

The models can hence be P(x|misfire), i.e. the distribution of the noise in a given variable,and P(x|fire), i.e. the expected value and its uncertainty. The advantage with this approach isthat the predictions can easily be adapted for an improved in-cycle detection performance, asis further explained herein.

The prior probability describes the expected risk of pilot pi1 misfire for a given operatingcondition before any additional information is available. lt can therefore be modelled as anormal distribution whose mean and standard deviation are functions of the start of injectionand the off-line calibrated pilot pi1 mass i.e. injector map. The uncertainty of the off-linecalibrated injectors map may be included in this model, which increases the total uncertaintyof pilot pi1 misfire. lf the injector 31 were individually calibrated on-line, the uncertainty couldbe reduced for better detection performance.

The ignition delay prediction has two applications. The first is to locate the range where tomeasure the magnitude of the pilot pi1 heat release Hr. This reduces the risk of confoundingoscillations with the actual heat release Hr magnitude. The second application is its use inthe pilot pi1 heat release Hr magnitude model. The ignition delay is related to the mixing ofthe fuel and hence how much is available at the time of auto-ignition. Rather than having ahighly accurate and complex model, a simple model to describe the position of the start ofthe pilot combustion C1 under a few crank angle degrees CAD can be sufficient.

The ignition delay can be used to estimate the position of the maximum pilot pi1 heat releaseHr, the accumulated heat release Hr and the pressure rise over the motoring. The magnitudeof the pilot pi1 heat release Hr can be modelled as a function of the injection on-time, railpressure and the in-cylinder pressure at the start of injection. The first two define the injectedpilot pi1 mass. The in-cylinder pressure can be correlated with the reactivity of the mixture.Additionally, the rail pressure has the effect of faster mixing and hence faster combustion,which increases the magnitude of the pilot pi1 heat release Hr. 22 The model can be empirical, and its uncertainty can be modelled as a normal distributionwith zero mean and a calibrated constant standard deviation. The model of the magnitude ofthe heat release Hr under misfire can be a model of the expected noise and oscillations ofthe heat release Hr. lt can hence be modelled as a normal distribution with a calibratedconstant mean and standard deviation.

The accumulated heat release Hr can be modelled as a linear function of the magnitude ofthe pilot pi1 heat release Hr. This simple model makes use of the fact that for small pilot pi1masses the combustion duration is substantially constant. Hence, the accumulated heatrelease Hr can be approximated as proportional to the heat release Hr magnitude. Theuncertainty can be modelled as a normal distribution with zero mean and a calibrated constant standard deviation.

Similar to the modelling of the magnitude of the heat release Hr, the model for theaccumulated heat release Hr subject to pilot pi1 misfire can be modelled as a normaldistribution, with zero mean and a calibrated constant standard deviation.

The pressure rise over the motoring pressure trace can be directly correlated with the energyreleased during the combustion. The benefit in using the pressure is that only the motoringpressure needs to be estimated, but no computation of the pressure derivative is required.The model, therefore, has the same structure as the model for the accumulated heat releaseHr. The model for the pressure rise subject to pilot pi1 misfire can be modelled as a normaldistribution, with zero mean and a calibrated constant standard deviation.

A wide spectrum of operating conditions can be required for the stochastic characterization.Mostly, the pilot pi1 mass and start of injection, but also engine speed and rail pressure canbe the significant variables to the misfire dispersion. On the other hand, for the predictivestochastic models, a suitable design of experiments with the variables used can be sufficient,which reduces the calibration effort despite the larger number of parameters.

The information from the misfire diagnosis once the cycle is completed can also be used toupdate the stochastic models. The adapted stochastic models can be used for the pilot pi1misfire detection of the following cycles, which will improve the performance and robustnessof the algorithm. Furthermore, more complex adaptation algorithms can be used for improvedadaptation properties, such as a non-linear Kalman filter. 23 According to some embodiments of the present disclosure, the control arrangement 9 isconfigured to perform a first separate diagnosis of the presence of a misfire in the currentcombustion cycle based on the at least first input value using one or more calculationvariables. Moreover, according to these embodiments, the control arrangement 9 may beconfigured to obtain at least a second input value indicative of the current combustion statein the current combustion cycle in the cylinder 3 of the engine 1. Furthermore, according tothese embodiments, the control arrangement 9 may be configured to perform a secondseparate diagnosis of the presence of a misfire in the current combustion cycle based on theat least second input value using one or more calculation variables. Furthermore, accordingto these embodiments, the control arrangement 9 may be configured to combine the first andsecond separate diagnoses into the diagnosis of the presence of a misfire in the currentcombustion cycle. Thereby, an even more robust and accurate diagnosis can be performed.This because at least two different input values are used to perform two separate diagnoseswhich are combined into the diagnosis of the presence of a misfire in the current combustioncycle. ln this manner, an even more accurate control of combustion parameters of theinternal combustion engine 1 can be performed during the current combustion cycle basedon the diagnosis.

According to some embodiments, the control arrangement 9 may be configured to evaluatethe first separate diagnosis by comparing the first separate diagnosis and the progress of theat least first input value during the current combustion cycle, and may be configured toprovide a first weighting factor based on the evaluation of the first separate diagnosis.Moreover, the control arrangement 9 may be configured to evaluate the second separatediagnosis by comparing the second separate diagnosis and the progress of the at leastsecond input value during the current combustion cycle, and to provide a second weightingfactor based on the evaluation of the second separate diagnosis. Furthermore, the controlarrangement 9 may be configured to, during the current combustion cycle, and/or during asubsequent combustion cycle, combine the first and second separate diagnoses into thediagnosis of the presence of a misfire using the first and second weighting factors. Thereby,an even more robust and accurate diagnosis can be performed. This because at least twodifferent input values are used to perform two separate diagnoses which are evaluated andcombined into the diagnosis of the presence of a misfire in the current combustion cycleusing the first and second weighting factors. ln this manner, an even more accurate controlof combustion parameters of the internal combustion engine 1 can be performed during thecurrent combustion cycle, and/or during a subsequent combustion cycle, based on thediagnosis. 24 The at least first input value may be indicative of one of a position of heat release Hr, acurrent heat release Hr, a current accumulated heat release Hr, and a current cylinderpressure. The at least second input value may be different from the at least first input valueand may be indicative of one of a position of heat release Hr, a current heat release Hr, a current accumulated heat release Hr, and a current cylinder pressure.

According to further embodiments, the control arrangement 9 may be configured to performmore than two separate diagnoses, such as three, four, five, or six separate diagnoses. Thecontrol arrangement 9 may be configured to combine such separate diagnoses into thediagnosis of the presence of a misfire in the current combustion cycle using a number ofweighting factors.

As understood from the above, several variables can be used as indicators for the pilot pi1misfire, the position of the pilot pi1 heat release Hr magnitude, the pilot pi1 heat release Hrmagnitude, the accumulated heat release Hr and the pressure rise over motoring. Each ofthese variables provides an indicator for the pilot pi1 misfire probability. Hence, by combiningadequately these indicators, the detection performance can be increased. lt can be desired to rely on those indicators that show the best detection performance. lf noprior information is known about how accurate each indicator is, only the uncertainty from thecalibration can be used to balance the fusion of the probability indicators. However, as donepreviously, the performance of each indicator can be evaluated on-line for an adaptivecombination of them. The proposed combination can be a weighted averaged of thepredicted probabilities of misfire and combustion. The weights can be updated proportionallyto the successful ratio. As before, the distance to the threshold can be used as a measure ofthe indicator robustness i.e. how far from the threshold each probability is. The weights canbe increased when the detection is successful and decreased when it is not. Finally, theweights can be normalized before computing the weighted average.

The control arrangement 9 may be configured to input one or more input values indicative ofa current operational state of the combustion engine 1 and to adapt the one or morecalculation variables based on the current operational state of the combustion engine 1.Such one or more input values may for example be indicative of a current rotational speed ofthe engine 1, a current load of the engine, a current rail pressure, a current flow of exhaustgas recirculation EGR, a current exhaust gas temperature, a current exhaust gascomposition, and the like.

Fig. 4 illustrates a vehicle 60, according to some embodiments. The vehicle 60 compriseswheels 62 and an engine 1 according to the embodiments illustrated in Fig. 1. The engine 1is configured to provide motive power to the vehicle 60 via one or more of the wheels 62 ofthe vehicle 60. The vehicle 60 illustrated in Fig. 4 is a truck. However, the engine 1, asreferred to herein, may be comprised in another type of manned or unmanned vehicle forland or water based propulsion such as a lorry, a bus, a construction vehicle, a tractor, a car,a boat, a ship or the like. Furthermore, the engine 1, as referred to herein, may be astationary internal combustion engine, for example an internal combustion engine of an engine driven generator.

Fig. 5 illustrates a method 100 of performing one or more in-cycle misfire diagnoses of thecombustion in a cylinder of an internal combustion engine. The engine 1 may be an engine 1according to the embodiments illustrated in Fig. 1 and Fig. 4 and some aspects are explainedwith reference to Fig. 2 and Fig. 3. Therefore, below, simultaneous reference is made to Fig.1 - Fig. 5.

The method 100 is a method 100 of performing one or more in-cycle misfire diagnoses of thecombustion in a cylinder 3 of an internal combustion engine 1. As indicated in Fig. 5, themethod 100 comprises the steps of: - obtaining 110 at least a first input value indicative of a current combustion statein a current combustion cycle in the cylinder 3, - performing 120 a diagnosis of the presence of a misfire during the currentcombustion cycle based on the at least first input value using one or morecalculation variables, - evaluating 130 the diagnosis by comparing the diagnosis and the progress ofthe at least first input value during the current combustion cycle, and - updating 140 the one or more calculation variables based on the evaluation ofthe diagnosis.

As indicated in Fig. 5, the step of performing 120 the diagnosis magi-comprises the step of: - performing 121 the diagnosis of the presence of a misfire during, and/or after, apilot injection pi1 and before the start of a time interval Ti in which a followingcombustion C2, C3 can occur as a result of a second pilot injection pi2, or a main injection mi, of fuel into the cylinder 3 in the current combustion cycle.

Moreover, as indicated in Fig. 5, the step of evaluating 130 the diagnosis -may--compriseg thestep of: 26 - evaluating 131 the diagnosis by comparing the diagnosis and the progress ofthe at least first input value during a time interval Ti in which a followingcombustion C2, C3 can occur as a result of a second pilot injection pi2, or a main injection mi, of fuel into the cylinder 3 in the current combustion cycle.

Furthermore, as indicated in Fig. 5, the step of updating 140 the one or more calculation variables may comprise the steps of: - determining 141 the accuracy of the diagnosis by comparing the diagnosis andthe progress of the at least first input value, and - updating 142 the one or more calculation variables if the accuracy of thediagnosis is below a threshold accuracy.

According to some embodiments, the one or more calculation variables is/are one or more thresholds, and wherein the step of performing 120 the diagnosis may comprise the step of: - performing 122 the diagnosis of the presence of a misfire by comparing the atleast first input value and the one or more thresholds.

As indicated in Fig. 5, the step of updating 140 the one or more calculation variables may comprise the step of: - updating 143 the one or more calculation variables based on the differencebetween the at least first input value and the one or more thresholds. l\/loreover, as indicated in Fig. 5, the step of updating 143 the one or more calculation variables may comprise the step of: - updating 144 the one or more calculation variables with a magnitude beingsubstantially inversely proportional to the difference between the at least firstinput value and the one or more thresholds.

According to some embodiments, the one or more calculation variables is/are one or more misfire probabilities.

As indicated in Fig. 5, the method 100 may comprise the step of:- obtaining 115 at least one misfire probability of the one or more misfireprobabilities by modelling the probability of a misfire using the at least first input value. l\/loreover, as indicated in Fig. 5, the method 100 may comprise the step of: 27 obtaining 116 at least one misfire probability of the one or more misfireprobabilities by computing the probability of a misfire with a stochastic modelusing the at least first input value.

Furthermore, as indicated in Fig. 5, the method 100 may comprise the step of: adapting 150 combustion parameters of the internal combustion engine 1during the current combustion cycle based on the diagnosis.

As indicated in Fig. 5, the step of performing 120 the diagnosis may comprise the steps of: performing 124 a first separate diagnosis of the presence of a misfire in thecurrent combustion cycle based on the at least first input value using one ormore calculation variables, obtaining 112 at least a second input value indicative of the current combustionstate in the current combustion cycle, performing 125 a second separate diagnosis of the presence of a misfire in thecurrent combustion cycle based on the at least second input value using one ormore calculation variables, and combining 126 the first and second separate diagnoses into the diagnosis ofthe presence of a misfire in the current combustion cycle. l\/loreover, as indicated in Fig. 5, the method 100 may comprise the steps of: evaluating 134 the first separate diagnosis by comparing the first separatediagnosis and the progress of the at least first input value during the currentcombustion cycle, providing 135 a first weighting factor based on the evaluation of the firstseparate diagnosis, evaluating 136 the second separate diagnosis by comparing the secondseparate diagnosis and the progress of the at least second input value duringthe current combustion cycle, and providing 137 a second weighting factor based on the evaluation of the secondseparate diagnosis, and wherein the method 100 may comprise the step of, during the current combustion cycle and/or during a subsequent combustion cycle: combining 139 the first and second separate diagnoses into the diagnosis ofthe presence of a misfire using the first and second weighting factors. 28 As mentioned herein, the at least first input value may be indicative of one of a position ofheat release Hr, a current heat release Hr, a current accumulated heat release Hr, and a current cylinder pressure.

Moreover, as mentioned herein, the at least second input value may different from the atleast first input value and may be indicative of one of a position of heat release Hr, a currentheat release Hr, a current accumulated heat release Hr, and a current cylinder pressure.

As indicated in Fig. 5, the method 100 may comprise: - inputting 102 one or more input values indicative of a current operational stateof the combustion engine 1, and - adapting 104 the one or more calculation variables based on the currentoperational state of the combustion engine 1. lt will be appreciated that the various embodiments described for the method 100 are allcombinable with the control arrangement 9 as described herein. That is, the controlarrangement 9 may be configured to perform any one of the method steps 102, 104, 110,112,115,116,120,121,122,124,125,126,130,131,134,135,136,137,139,140,141,142, 143, 144, and 150 of the method 100.

Fig. 6 illustrates computer-readable medium 200, according to some embodiments. Thecomputer-readable medium 200 comprises instructions which, when executed by acomputer, cause the computer to carry out the method 100 according to some embodimentsdescribed herein.

According to some embodiments, the computer-readable medium 200 comprises a computerprogram comprising instructions which, when the program is executed by a computer, causethe computer to carry out the method 100 according to some embodiments.

The control arrangement 9, as referred to herein, may be connected to one or morecomponents of the engine 1, and to one or more components of a vehicle comprising theengine, in order to perform the method 100 illustrated in Fig. 5.

One skilled in the art will appreciate that the method 100 of performing one or more in-cyclemisfire diagnoses of the combustion in a cylinder 3 of an internal combustion engine 1 maybe implemented by programmed instructions. These programmed instructions are typicallyconstituted by a computer program, which, when it is executed in the control arrangement 9, 29 ensures that the control arrangement 9 carries out the desired control, such as the methodsteps102,104,110,112,115,116,120,121,122,124,125,126,130,131,134,135,136,137, 139, 140, 141, 142, 143, 144, and 150 described herein. The computer program isusually part of a computer program product 200 which comprises a suitable digital storagemedium on which the computer program is stored.

The control arrangement 9 may comprise a calculation unit which may take the form ofsubstantially any suitable type of processor circuit or microcomputer, e.g. a circuit for digitalsignal processing (digital signal processor, DSP), a Central Processing Unit (CPU), aprocessing unit, a processing circuit, a processor, an Application Specific Integrated Circuit(ASIC), a microprocessor, or other processing logic that may interpret and executeinstructions. The herein utilised expression “calculation unit” may represent a processingcircuitry comprising a plurality of processing circuits, such as, e.g., any, some or all of the ones mentioned above.

The control arrangement 9 may further comprise a memory unit, wherein the calculation unitmay be connected to the memory unit, which may provide the calculation unit with, forexample, stored program code and/or stored data which the calculation unit may need toenable it to do calculations. The calculation unit may also be adapted to store partial or finalresults of calculations in the memory unit. The memory unit may comprise a physical deviceutilised to store data or programs, i.e., sequences of instructions, on a temporary orpermanent basis. According to some embodiments, the memory unit may compriseintegrated circuits comprising silicon-based transistors. The memory unit may comprise e.g.a memory card, a flash memory, a USB memory, a hard disc, or another similar volatile ornon-volatile storage unit for storing data such as e.g. ROIVI (Read-Only l\/lemory), PROIVI(Programmable Read-Only Memory), EPROIVI (Erasable PROIVI), EEPROIVI (ElectricallyErasable PROIVI), etc. in different embodiments.

The control arrangement 9 is connected to components engine 1 for receiving and/orsending input and output signals. These input and output signals may comprise waveforms,pulses or other attributes which the input signal receiving devices can detect as informationand which can be converted to signals processable by the control arrangement 9. Thesesignals may then be supplied to the calculation unit. One or more output signal sendingdevices may be arranged to convert calculation results from the calculation unit to outputsignals for conveying to other parts of the vehicle's control system and/or the component orcomponents for which the signals are intended. Each of the connections to the respectivecomponents of the engine 1 for receiving and sending input and output signals may take the form of one or more from among a cable, a data bus, e.g. a CAN (controller area network)bus, a MOST (media orientated systems transport) bus or some other bus configuration, or a wireless connection. ln the embodiments illustrated, the engine 1 comprises a control arrangement 9 but mightalternatively be implemented wholly or partly in two or more control arrangements or two or more control units.

Control systems in modern vehicles generally comprise a communication bus systemconsisting of one or more communication buses for connecting a number of electronic controlunits (ECUs), or controllers, to various components on board the vehicle. Such a controlsystem may comprise a large number of control units and taking care of a specific functionmay be shared between two or more of them. Vehicles of the type here concerned aretherefore often provided with significantly more control arrangements than depicted in Fig. 1as one skilled in the art will surely appreciate.

The computer program product 200 may be provided for instance in the form of a data carriercarrying computer program code for performing at least some of the method steps 102, 104,110,112,115,116,120,121,122,124,125,126,130,131,134,135,136,137,139,140,141, 142, 143, 144, and 150 according to some embodiments when being loaded into one ormore calculation units of the control arrangement 9. The data carrier may be, e.g. a CD ROIVIdisc, as is illustrated in Fig. 6, or a ROIVI (read-only memory), a PROIVI (programable read-only memory), an EPROIVI (erasable PROIVI), a flash memory, an EEPROIVI (electricallyerasable PROIVI), a hard disc, a memory stick, an optical storage device, a magnetic storagedevice or any other appropriate medium such as a disk or tape that may hold machinereadable data in a non-transitory manner. The computer program product may furthermorebe provided as computer program code on a server and may be downloaded to the controlarrangement 9 remotely, e.g., over an Internet or an intranet connection, or via other wired or wireless communication systems. lt is to be understood that the foregoing is illustrative of various example embodiments andthat the invention is defined only by the appended claims. A person skilled in the art willrealize that the example embodiments may be modified, and that different features of theexample embodiments may be combined to create embodiments other than those describedherein, without departing from the scope of the present invention, as defined by theappended claims. 31 As used herein, the term "comprising" or "comprises" is open-ended, and includes one ormore stated features, elements, steps, components, or functions but does not preclude thepresence or addition of one or more other features, elements, steps, components, functions,or groups thereof.

The wording “in-cycle misfire diagnosis”, as used herein, means that the diagnosis isperformed in the current ongoing combustion cycle. The wording “current combustion cycle”as used herein may be defined as the same ongoing combustion cycle.

Claims

CLA|l\/IS

1. A method (100) of performing one or more in-cycle misfire diagnoses of the combustion in a cylinder (3) of an internal combustion engine (1), wherein the method (100) comprises: obtaining (110) at least a first input value indicative of a current combustion state in acurrent combustion cycle in the cylinder (3), performing (120. tät) a diagnosis of the presence of a misfire during andior after a ' the current combustion cycle and before the start of a tline intervai (Ti) in tvltlch a tclicwlnci combustion ((32, 12333 can cccur as a result nt asecond pilot inlectlon (plåt, cr a main lniecticiw (mi), of fuel into the cvilnder (3) in thecurrent ccrnbtsstlcrt cycle. based on the at least first input value using one or morecalculation variables, evaluating (130gl§j_) the diagnosis by comparing the diagnosis and the progress of the at least first input value during a tirne lntertfai (Ti) af the current combustion cycle; ln tfrlticri time inierval (Ti) a tolicvvinc combustion (G2, (33) can cccur as a resuii of asecond pilot inlecllcn (m2) or a rriain litiecticn (rnil ct fuel into the ctfilntzler (3) in thecurrent carntz-ustiort cycle, and updating (140) the one or more calculation variables based on the evaluation of the diagnosis. \\\\\\\\\\\\\\\ “The method (100) according to claima_l_, wherein the step of updating (140) the one or more calculation variables comprises the steps of: 2 - determining (141) the accuracy of the diagnosis by comparing the diagnosis and theprogress of the at least first input value, and - updating (142) the one or more calculation variables if the accuracy of the diagnosisis below a threshold accuracy. \\\\\\\\\\\\\\\\\\\\\ “The method (100) according to any one of the preceding claims, wherein theone or more calculation variables is/are one or more thresholds, and wherein the step ofperforming (120) the diagnosis comprises the step of: - performing (122) the diagnosis of the presence of a misfire by comparing the at leastfirst input value and the one or more thresholds. method (100) according to claim wherein the step of updating (140) theone or more calculation variables comprises the step of: - updating (143) the one or more calculation variables based on the differencebetween the at least first input value and the one or more thresholds. _____________________ “The method (100) according to claim fêfå, wherein the step of updating (143) theone or more calculation variables comprises the step of:- updating (144) the one or more calculation variables with a magnitude beingsubstantially inversely proportional to the difference between the at least first input value and the one or more thresholds. \\\\\\\\\\\\\\\\\\\\\ “The method (100) according to any one of the preceding claims, wherein the one or more calculation variables is/are one or more misfire probabilities. method (100) according to claim êﬁ, wherein the method (100) comprisesthe step of:- obtaining (115) at least one misfire probability of the one or more misfire probabilitiesby modelling the probability of a misfire using the at least first input value. \\\\\\\\\\\\\\\\\ “The method (100) according to claim gå or 92, wherein the method (100)comprises the step of:- obtaining (116) at least one misfire probability of the one or more misfire probabilitiesby computing the probability of a misfire with a stochastic model using the at least first input value. 3 "f \\\\\\\\\\\\\\\\\ “The method (100) according to any one of the preceding claims, wherein the _________ ~\ method (100) comprises the step of: adapting (150) combustion parameters of the internal combustion engine (1) duringthe current combustion cycle based on the diagnosis. \\\\\\\\\\\\\\\ “The method (100) according to any one of the preceding claims, wherein the step of performing (120) the diagnosis comprises the steps of: performing (124) a first separate diagnosis of the presence of a misfire in the currentcombustion cycle based on the at least first input value using one or more calculationvariables, obtaining (112) at least a second input value indicative of the current combustionstate in the current combustion cycle, performing (125) a second separate diagnosis of the presence of a misfire in thecurrent combustion cycle based on the at least second input value using one or morecalculation variables, and combining (126) the first and second separate diagnoses into the diagnosis of the presence of a misfire in the current combustion cycle. ggggggggggggggg “The method (100) according to claim -f-åíg, wherein the method (100) comprises the steps of: evaluating (134) the first separate diagnosis by comparing the first separatediagnosis and the progress of the at least first input value during the currentcombustion cycle, providing (135) a first weighting factor based on the evaluation of the first separatediagnosis, evaluating (136) the second separate diagnosis by comparing the second separatediagnosis and the progress of the at least second input value during the currentcombustion cycle, and providing (137) a second weighting factor based on the evaluation of the secondseparate diagnosis, and wherein the method (100) comprises the step of, during the current combustion cycle and/or during a subsequent combustion cycle: combining (139) the first and second separate diagnoses into the diagnosis of thepresence of a misfire using the first and second weighting factors. if: v*“Pèt - .« . 4 \\\\\\\\\\\\\\\\\\\\\\ “The method (100) according to any one of the preceding claims, wherein the at least first input value is indicative of one of a position of heat release (Hr), a current heatrelease (Hr), a current accumulated heat release (Hr), and a current cylinder pressure. ___________The method (100) according to claim Il-S-jgnor 1311, wherein the at leastsecond input value is different from the at least first input value and is indicative of one ofa position of heat release (Hr), a current heat release (Hr), a current accumulated heatrelease (Hr), and a current cylinder pressure. \\\\\\\\\\\\\\\\\\\\\\ “The method (100) according to any one of the preceding claims, wherein themethod (100) comprises:- inputting (102) one or more input values indicative of a current operational state ofthe combustion engine (1), and- adapting (104) the one or more calculation variables based on the current operational state of the combustion engine (1). ____________ “A computer program comprising instructions which, when the program isexecuted by a computer, cause the computer to carry out the method (100) according toany one of the claims 1 - jg-t-ë. A computer-readable medium (200) comprising instructions which, when executed by a computer, cause the computer to carry out the method (100) according toany one of the claims 1 - låt-å. control arrangement (9) for an internal combustion engine (1), wherein the control arrangement (9) is configured to perform one or more in-cycle misfire diagnoses of the combustion in a cylinder (3) of the internal combustion engine (1), wherein thecontrol arrangement (9) is configured to: - obtain at least a first input value indicative of a current combustion state in a currentcombustion cycle in the cylinder (3), - perform a diagnosis of the presence of a misfire during. and/or after. a iïiiiot inieotioslgigiutjin the current combustion cycle and belore lite start of a tirne irilerxfai 'Tit inxfvhlciw a tollovifinq combustion (G

2. C31 can occur as a result of a second pilotinåt-action (blå). or a main inlectloit imi), of fuel into the cylinder få) in the currentcombustion cyoie. based on the at least first input value using one or morecalculation variables, 5 - evaluate the diagnosis by comparing the diagnosis and the progress of the at leastfirst input value during a time interval til; of the current combustion cycle, in tflrtticlttime interval (lit a lotlowino combustion (C32 C33) cart occur ae a result ot a secondpilot iniectioit íošâl. or a enašn šnieetion (ntifi, of fueš into the cvšindee* (3Éi in the current 5 conttz-ustieit cycle. and - update the one or more calculation variables based on the evaluation of the diagnosis. internal combustion engine (1) comprising a control arrangement (9)| 10 according to claim iZi-Q. vehicle (60) comprising an internal combustion engine (1) according to claim