JP2009293501A - Misfire detecting device for multicylinder internal combustion engine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a misfire detecting device for a multicylinder internal combustion device capable of accurately detecting misfire even if there is mechanical dispersion of a crank angle sensor or a swing back of rotational speed following misfire. <P>SOLUTION: A first fluctuation amount Δω1<SB>n</SB>being a fluctuation value of a difference of rotational speeds per cylinder between cylinders wherein cycles are shifted for just an integral multiple of one rotation of an output shaft of the internal combustion engine, is calculated as Δω1<SB>n</SB>=(ω<SB>n</SB>-ω<SB>n-2</SB>)-(ω<SB>n-1</SB>-ω<SB>n-3</SB>) in for example, a four cylinder internal combustion engine, and a second fluctuation amount Δω2<SB>n</SB>being a fluctuation amount of the first fluctuation amount Δω1<SB>n</SB>is calculated as Δω2<SB>n</SB>=Δω1<SB>n</SB>-Δω1<SB>n-1</SB>. If the second fluctuation amount Δω2<SB>n</SB>is smaller than a determination value -A, and a second fluctuation amount Δω2<SB>n-2</SB>being a value prior to the last value is larger than the determination value A, misfire occurrence is determined. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a misfire detection device for detecting misfire in a multi-cylinder internal combustion engine.

In Patent Document 1, the crank angular speed fluctuation amount for each cylinder is obtained based on the rotation signal NE corresponding to the rotation of the crankshaft, and the misfire occurrence is detected by comparing the crank angular speed fluctuation amount for each cylinder with a predetermined misfire determination value. More specifically, a misfire detection device is disclosed, and more specifically, a difference in the crank angular speed fluctuation amount for each cylinder of a cylinder separated by an integer multiple of a crank angle of a crank angle of 720 deg or 360 deg or a stroke phase difference between cylinders is calculated, The occurrence of misfire is detected by individually comparing the difference calculation results with the misfire determination value.
JP-A-10-054295

By the way, when the misfire occurrence is detected based on the rotation signal NE corresponding to the rotation of the crankshaft as described above, the rotational speed swings back along with the misfire. It is desired to detect with high accuracy.
However, in the conventional misfire detection device, misfire may be misdiagnosed in the low rotation speed region due to the reversal of the rotational speed that occurs with the misfire.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a misfire detection device for a multi-cylinder internal combustion engine that can detect misfire with high accuracy even when the rotational speed is swung back due to misfire. .

  For this reason, the misfire detection apparatus for a multi-cylinder internal combustion engine according to the present invention detects a change in the rotational speed of each cylinder between cylinders whose cycles are shifted by an integral multiple of one rotation of the output shaft of the multi-cylinder internal combustion engine. The variation amount of the first variation amount is calculated as the first variation amount, the variation amount of the first variation amount is calculated as the second variation amount, and the presence or absence of misfire is determined based on the second variation amount. .

According to the above invention, if a change in the rotation speed of each cylinder between cylinders whose cycles are shifted by an integral multiple of one rotation of the output shaft of the multi-cylinder internal combustion engine is detected, the change is caused by a difference in torque generated by each cylinder ( The difference in the rotational speed for each cylinder due to the presence or absence of combustion) is indicated with high accuracy.
Further, when the variation amount of the change is calculated as the first variation amount, and further, the variation amount of the first variation amount is calculated as the second variation amount, the second variation amount emphasizes the change of the first variation amount. This makes it possible to detect misfire by distinguishing it from swinging back the rotational speed due to misfire.

  Therefore, the occurrence of misfire can be detected with high accuracy even if the rotational speed swings back due to misfire.

Embodiments of the present invention will be described below.
FIG. 1 is a system diagram of a vehicular multi-cylinder internal combustion engine to which a misfire detection apparatus according to the present invention is applied.
An electronic control throttle device 103 is interposed in the intake pipe 102 of the internal combustion engine 101 shown in FIG.

The air that has passed through the electronically controlled throttle device 103 is taken into the combustion chamber 105 through the intake valve 104.
Gas (exhaust gas) after combustion in the combustion chamber 105 is discharged to an exhaust pipe 107 through an exhaust valve 106 and purified by a front catalytic converter 108 and a rear catalytic converter 109 interposed in the exhaust pipe 107. And then released into the atmosphere.

The intake valve 104 and the exhaust valve 106 are opened and closed by an intake camshaft 110A and an exhaust camshaft 110B, respectively.
A fuel injection valve 112 is provided in the intake port 111 upstream of the intake valve 104 of each cylinder.
The fuel injection valve 112 injects fuel toward the intake valve 104 when driven to open by an injection pulse signal output from an engine control unit (hereinafter referred to as ECU) 113.

The fuel injection valve 112 may be an in-cylinder direct injection internal combustion engine in which fuel is directly injected into the combustion chamber 105.
The fuel sucked into the combustion chamber 105 together with air is ignited and burned by spark ignition by the spark plug 114.
Each spark plug 114 is provided with an ignition coil 115 having a built-in power transistor, and the ECU 113 controls on / off of the power transistor to start energization start / energization timing of the ignition coil 115. And the ignition timing (ignition advance value) of each cylinder is independently controlled.

The ECU 113 includes an accelerator pedal sensor 116 for detecting the accelerator opening APS, an air flow sensor 115 for detecting the intake air flow rate QA of the internal combustion engine 101, and a detected part such as a protrusion formed on a signal plate attached to the drive plate. Is detected, the crank angle sensor 117 that outputs a position signal POS every time the crankshaft 121 (output shaft) rotates by a unit crank angle (for example, 5 to 10 deg), and the throttle that detects the opening TVO of the throttle valve 103b. A sensor 118, a water temperature sensor 119 for detecting the coolant temperature Tw of the internal combustion engine 101, a cam sensor 120 for outputting a cylinder discrimination signal PHASE in synchronization with the rotation of the intake camshaft 110A, and an air-fuel ratio AF ( Air-fuel ratio sensor for detecting exhaust air-fuel ratio) In the crank angle sensor 117, the crank angle is positioned at two positions 180 ° apart from each other at a cycle several times (for example, three times) the angle corresponding to the unit crank angle. The detected portion is provided so that the signal POS is output, and the ECU 113 detects the reference crank angle position REF for every crank angle of 180 deg by detecting a portion where the output cycle of the position signal POS is long. To do.

The internal combustion engine 101 is an in-line four cylinder, and the reference crank angle position REF is detected every 180 deg which is a stroke phase difference (fuel injection interval / ignition interval) between the cylinders. Based on the position REF, fuel injection timing, ignition timing, etc. in each cylinder are detected.
Further, the ECU 113 measures the period of the reference crank angle position REF and calculates the rotational speed NE (rpm) of the internal combustion engine 101.

A reference crank angle sensor that outputs a reference crank angle signal REF every 180 degrees of crank angle may be provided together with a crank angle sensor 117 that outputs the position signal POS. In this case, the output of the position signal POS All the cycles can be made uniform in one rotation of the crankshaft 121.
The ECU 113 has a function of detecting misfire based on the engine rotational speed NE in software, and the misfire detection function will be described in detail with reference to the flowchart of FIG.

The routine shown in the flowchart of FIG. 2 is interrupted and executed every detection cycle (every timing at which the angular velocity is detected) of the rotation speed by cylinder (angular velocity by cylinder).
First, in step S1001, the angular velocity ω _n that is affected by the expansion stroke of each cylinder is detected for each combustion (for each cylinder) (rotational speed detecting means for each cylinder).
The angular speed ω _n (rotational speed per cylinder) is calculated by detecting a time T required to rotate by 60 deg from 110 deg (ATDC 110 deg) after top dead center to 170 deg (ATDC 170 deg) after top dead center. Is done.

However, the angular region for obtaining the angular velocity ω _n (rotational speed for each cylinder) is not limited to ATDC 110 deg to ATDC 170 deg for each cylinder, but is appropriately set as an optimal angular region for detecting changes in angular velocity due to misfire. .
In the four-cylinder engine 101 of this embodiment, one cycle (exhaust stroke → intake stroke → compression stroke → expansion stroke) of each cylinder is completed by two rotations of the crankshaft 121, and the ignition order is # 1 → # 3 → # 2. → # 4, and the strokes (cycles) of the cylinders are shifted from each other by 180 deg. When the compression stroke of the # 1 cylinder (piston bottom dead center → piston top dead center) ends, the compression stroke of the # 3 cylinder When the # 3 cylinder compression stroke ends, the # 2 cylinder compression stroke starts. When the # 2 cylinder compression stroke ends, the # 4 cylinder compression stroke starts.

Accordingly, each stroke of the # 1 cylinder and each stroke of the # 2 cylinder are shifted by one rotation (360 deg) of the crankshaft (output shaft) 121. Similarly, each stroke of the # 3 cylinder and # 4 cylinder These strokes are deviated by one rotation (360 deg) of the crankshaft 121.
In the signal plate of the crank angle sensor 117, an angle region A corresponding to 110 deg to 170 deg after the compression top dead center of the # 1 cylinder and an angle corresponding to 110 deg to 170 deg after the compression top dead center of the # 2 cylinder. This corresponds to the region A, and similarly corresponds to the angle region B corresponding to 110 deg to 170 deg after the compression top dead center of the # 3 cylinder and 110 deg to 170 deg after the compression top dead center of the # 4 cylinder. This corresponds to the angle region B (see FIG. 3).

The detection of the angular velocity ω _n for each cylinder in step S1001 is performed in detail as shown in the flowcharts of FIGS.
First, when the position of 110 deg after top dead center of each cylinder is detected based on the position signal POS from the crank angle sensor 117, the routine shown in the flowchart of FIG. 4 is interrupted and executed in step S1101. The value of the free run counter is stored as a value at 110 deg after top dead center.

On the other hand, when the position of 170 deg after the top dead center of each cylinder is detected based on the position signal POS from the crank angle sensor 117, the routine shown in the flowchart of FIG.
In step S1201, the value of the free-run counter at that time is stored as a value at 170 deg after top dead center.

In the next step S1202, as a difference between the value of the free run counter at 110 deg after the top dead center and the value of the free run counter at 170 deg after the top dead center, the crank angle is 60 deg from the time of 110 deg after the top dead center. The time T required to rotate and rotate to 170 deg after top dead center is calculated.
In step S1203, the angular velocity ω _n is calculated from the time T required to rotate by 60 deg at the crank angle.

A plurality of calculation results of the angular velocity ω _n are stored in the calculation order (in time series), the latest calculation result is ω _n , and the previous calculation result (before 180 deg) is ω _n−1 . There, the calculation result of the time before last (360 deg ago) is omega _n-2, calculation results than the omega _n-2 180 deg before (540Deg ago) is omega _n-3, a new angular velocity omega _n is calculated each that it latest value omega _n up is stored as the previous value omega _n-1, the previous value omega _n-1 until it is adapted to be updated and stored as the second preceding value omega _n-2.

When the angular velocity ω _n is calculated in step S1001, in step S1002 (detection means and first fluctuation amount calculation means), the first fluctuation amount Δω1 _n is calculated according to the following equation.
“Δω1 _n = (ω _n −ω _n-2 ) − (ω _n−1 −ω _n-3 )”
In the above equation, the difference “ω _n −ω _n−2 ” is the difference between the latest value ω _n and the value ω _n−2 before 360 deg. Since both are calculated based on the result of measuring the period of the same angle range of the signal plate of the crank angle sensor 117, for example, ω _n due to mechanical variation of the crank angle sensor 117 (signal plate). And ω _n−2 , even if the angle is deviated from 60 °, the influence is equally added to ω _n and ω _n−2 .

Therefore, the difference “ω _n −ω _n−2 ” eliminates the influence of mechanical variation of the crank angle sensor 117 (signal plate), and indicates the difference in actual angular velocity ω _n with high accuracy.
Similarly, the difference “ω _n-1 −ω _n-3 ” is the difference between the value ω _n-1 before 180 deg and the value ω _n-3 before 540 deg, and is shifted by one rotation of the crankshaft 121. Both are the results of measuring the elapsed time in the same angular range of the signal plate of the crank angle sensor 117.

Therefore, the difference “ω _n-1 −ω _n-3 ” also eliminates the influence of mechanical variations of the crank angle sensor 117 (signal plate), and shows the difference in actual angular velocity with high accuracy.
The first fluctuation amount Δω1 _n indicates the fluctuation amount of the difference in angular velocity (rotational speed for each cylinder) between the cylinders shifted by one rotation of the crankshaft 121. As described above, each difference represents the crank angle sensor. Since the influence of the mechanical variation of 117 (signal plate) is eliminated, the first variation Δω1 _n is also a value that is not affected by the mechanical variation of the crank angle sensor 117 (signal plate).

In step S1002, as described above, the difference in the rotational speed of each cylinder between the cylinders whose cycles are shifted by an integral multiple of one rotation of the crankshaft 121 (output shaft) is calculated, and the variation amount of the difference is determined as the first variation. It is calculated as an amount Δω1 _n , in other words, a change in the rotational speed of each cylinder between cylinders whose cycle is shifted by an integral multiple of one rotation of the crankshaft 121 (output shaft) is detected, and the amount of change in the change Is calculated as the first fluctuation amount Δω1 _n .

After calculating the first variation amount .DELTA..omega.1 _n, the next step S1003 (second variation calculation means), according to the following equation to calculate a second variation amount ?? 2 _n is the amount of variation in the first variation amount .DELTA..omega.1 _{n.
"Δω2 n = Δω1 n -Δω1 n-} 1 "
In the above equation, the first fluctuation amount Δω1 _n is a value calculated in step S1002 this time, and the first fluctuation amount Δω1 _n-1 is a value calculated when the processing in step S1002 was executed last time (before 180 deg). is there.

In step S1004, it is determined whether or not the second fluctuation amount Δω2 _n calculated in step S1003 is smaller than the determination value −A (<0) (whether Δω2 _n <−A or Δω2 _n ≦ −A). Judge whether or not).
If the second fluctuation amount Δω2 _n is a negative value and the absolute value is larger than A, the process proceeds to step S1005.

The determination value -A can be a constant value stored in advance, or can be variably set according to the engine operating conditions (for example, engine load and / or engine speed).
The second fluctuation amount Δω2 _n becomes a value close to zero when there is no misfire, and when the misfire occurs, the second fluctuation amount Δω2 _n swings in a minus direction and a plus direction. Since it varies depending on conditions, the misfire can be detected with higher accuracy by increasing the absolute value of the determination value -A as the operating condition increases the amplitude of the second fluctuation amount Δω2 _n associated with the occurrence of misfire. become.

Further, instead of variably setting the judgment value −A according to the engine operating conditions (for example, engine load and / or engine speed), the difference “ω _n −ω _n−2 ”, “ω _n−1 −ω _n-3 ”, at least one of the first variation Δω1 _n and the second variation Δω2 _n is corrected according to the engine operating condition (for example, the engine load and / or the engine speed), and a predetermined determination value − A can be compared with the corrected second variation Δω2 _n .

In step S1005, the second preceding value (360 deg previous value) of the second variation amount ?? 2 _n ?? 2 _n-2 is, the determination value A (> 0) greater determines whether than (?? 2 _n> A or ?? 2 _n It is determined whether or not ≧ A).
When the second fluctuation amount Δω2 _n−2 is a positive value and its absolute value is larger than A, that is, the second fluctuation amount Δω2 _n is smaller than the determination value −A, and the second fluctuation amount Δω2 _{If n-2} is larger than the determination value A, the process proceeds to step S1006 to determine the occurrence of misfire, and further, the cylinder that was in the expansion stroke (explosion stroke) at the timing at which Δω2 _n-3 is calculated is set to the misfire cylinder. As specified.

However, it can be configured such that only the occurrence of misfire is determined and the misfire cylinder is not specified.
The above-described steps S1004 to S1006 indicate the function as misfire determination means.
Here, the determination value A is a positive value having the same absolute value as the determination value -A. If the determination value -A is a constant value, the determination value A is a constant value, and the determination value -A is in an engine operating state. If the configuration is variably set in response, the setting is similarly variably set in accordance with the engine operating state.

Further, the determination value to be compared with the second variation amount ?? 2 _n (<0), the second variation amount ?? 2 _n-2 and compared to determine value (> 0) can be set to different values of the absolute value .
Here, as shown in FIG. 6, the first fluctuation amount Δω1 _n changes to the minus side due to the occurrence of misfire. However, the first fluctuation amount Δω1 _n is substantially equal to the level based on misfire even when the rotational speed is swung back due to misfire. It shows a negative value of the same level.

For this reason, if it is set as the structure which determines the presence or absence of misfire based on whether the 1st variation | change_quantity ₍ DELTA) ₍ omega) 1n fell below the negative judgment value, it cannot distinguish between misfire and the rocking | fluctuation accompanying misfire, and misfire generation May be erroneously determined.
In contrast, the second variation amount ?? 2 _n, since illustrates the variation of the first variation amount .DELTA..omega.1 _n, if the first variation amount .DELTA..omega.1 _n is changed from negative to positive across zero, and the first When the fluctuation amount Δω1 _n changes from plus to minus across zero, the change is emphasized and shows a larger change range.

Accordingly, the rotational speed decreases due to misfire, and the reaction speed greatly increases and the rotational speed again returns to the vicinity of the steady level, so that the second fluctuation amount Δω2 _n is clearly changed in the positive direction and immediately after the negative direction. In order to determine the change in the positive direction and the change in the negative direction immediately thereafter, misfire can be detected by matching the determination values A and -A.

On the other hand, when the rotational speed rises due to a reaction caused by misfire and the rotational speed returns to the vicinity of the steady level again, and further exceeds the steady level, and the rotational speed further decreases, the rotational speed further decreases from near the steady level, so the first fluctuation amount Δω1. The change amount of _n is relatively small, and the change in the negative direction of the second fluctuation amount Δω2 _n is a change when the rotation speed decreases to near the steady level again after a large increase in rotation due to the reaction of misfire. The amount is clearly different.

Therefore, if misfire detection is performed on the basis of the second variation Δω2 _n , the misfire is erroneously detected without being affected by the mechanical variation of the crank angle sensor 117 and by swinging back due to misfire. The occurrence of misfire can be detected with high accuracy.
If misfiring caused by misfire is misjudged as misfiring, a cylinder that has not actually misfired is judged as a misfiring cylinder, so that a useless failure such as stopping fuel injection to a normally burning cylinder When safe processing is performed or the misfire frequency is excessively calculated, engine operation is unduly restricted, and further, diagnostic information different from the actual situation is provided to the driver.

  On the other hand, in the above-described embodiment, misfire can be distinguished from shakeback due to misfire, and even if a shakeback occurs, a misfired cylinder is not erroneously determined based on this, so a cylinder that is burning normally Therefore, the fail-safe process for the misfired cylinder is not applied, the operation of the engine is not unduly restricted, and more reliable diagnostic information can be provided to the driver. Become.

By the way, the misfire detection based only on the second variation Δω2 _n is relatively small in the variation in combustion among the cylinders and is effective when misfires occur irregularly. If it becomes larger or the single cylinder misfires each time, it becomes difficult to distinguish the misfire swingback, and the misfire determination accuracy may be lowered.
Accordingly, a second embodiment will be described in accordance with the flowchart of FIG. 7 in which the combustion variation among the cylinders is relatively large and the detection accuracy can be maintained even if the single cylinder misfires each time.

In the flowchart of FIG. 7, in steps S2001 to S2003, the angular velocity ω _n , the first variation amount Δω1 _n , and the second variation amount Δω2 _n are calculated in the same manner as in steps S1001 to S1003.
In step S2004, as in step S1004, it is determined whether or not the second variation amount Δω2 _n is smaller than the determination value −A (determines whether Δω2 _n <−A or Δω2 _n ≦ −A. ).

Note that the determination value -A is set in the same manner as in step S1004.
_{If it} is determined in step S2004 that the second variation amount Δω2 _n is smaller than the determination value −A, the process proceeds to step S2005, and whether or not the first variation amount Δω1 _n−1 is larger than the determination value B (> 0). (Determining whether Δω1 _n-1 > B or Δω1 _n-1 ≧ B).

The determination value B can be set to a constant value similarly to the determination value -A, and can be variably set according to engine operating conditions (for example, engine load and / or engine speed).
The magnitude relationship between the absolute value of the determination value A and the absolute value of the determination value B should be specified according to the type of the internal combustion engine and the operating conditions, and is not uniquely determined.
If it is determined in step S2005 that the first fluctuation amount Δω1 _n-1 is larger than the determination value B, the process proceeds to step S2006, and the second fluctuation amount Δω2 _n-2 is larger than the determination value A as in step S1005. It is determined whether or not it is large (determines whether or not Δω2 _n > A or Δω2 _n ≧ A).

The determination value A is set in the same manner as in step S1005.
If it is determined in step S2006 that the second variation amount Δω2 _n−2 is larger than the determination value A, the process proceeds to step S2007, and whether the first variation amount Δω1 _n−3 is smaller than the determination value −B (<0). (Δω1 _n−3 <−B or Δω1 _n−3 ≦ −B is determined).

The relationship between the determination value B and the determination value -B is the same as the relationship between the determination value A and the determination value -A.
That is, the determination value B and the determination value −B are constant values or values that are variably set according to the engine operating conditions, and the determination value B and the determination value −B are different in absolute value. be able to.
When it is determined in step S2007 that the first fluctuation amount Δω1 _n-3 is smaller than the determination value −B, that is, Δω2 _n <−A or Δω2 _n ≦ −A, and Δω1 _n-1 > B or Δω1. _{If n−1} ≧ B, Δω2 _n > A or Δω2 _n ≧ A, and Δω1 _n-3 <−B or Δω1 _n-3 ≦ −B, the process proceeds to step S2008, and misfire occurs. Further, the cylinder that was in the expansion stroke (explosion stroke) at the timing at which Δω1 _n-3 (Δω2 _n-3 ) was calculated is specified as the misfiring cylinder.

In 2nd Embodiment, the part of step S2004-step S2008 shows the function as a misfire determination means.
As described above, in the second embodiment, misfire detection is performed from the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n, and the condition for misfire determination in the first embodiment is Δω1 _n−1 > B. Or Δω1 _n−1 ≧ B and Δω1 _n-3 <−B or Δω1 _n-3 ≦ −B, and even in a state where misfire determination is made in the first embodiment. Furthermore, the misfire determination is not made unless the first variation Δω1 satisfies a predetermined condition.

Among the added determination conditions, Δω1 _n-3 <−B or Δω1 _n-3 ≦ −B is used to determine a drop in the angular velocity ω _n due to misfire, and Δω1 _n-1 > B or Δω1 _n-1 ≧ B is used to determine an increase in the angular velocity ω _n due to the rotation swinging back due to misfire.
That is, when both the first variation amount Δω1 _n and the second variation amount Δω2 _n indicate a change pattern at the time of misfiring, the misfire determination is performed, whereby the second variation amount Δω2 _n is changed between the cylinders. Even if a change pattern similar to that at the time of misfire is shown due to the variation in combustion, misfire can be prevented from being erroneously detected.

Conversely, even if the first fluctuation amount Δω1 _n is influenced by combustion variations and shows a change pattern at the time of misfire, if the second fluctuation amount Δω2 _n does not change along the misfire pattern, misfire is detected erroneously. It will not be done.
When the first variation amount Δω1 _n changes across zero, the second variation amount Δω2 _n becomes a value that emphasizes the change. For example, Δω1 _n−1 becomes Δω1 _n−1 > B or Δω1 _{n. Even if} it is a positive value that does not satisfy the condition of ₋₁ ≧ B, if Δω1 _n greatly changes to the negative side, the condition of Δω2 _n <−A or Δω2 _n ≦ −A may be satisfied. There is.

Similarly, even if Δω1 _n-3 is a negative value that does not satisfy Δω1 _n-3 <−B or Δω1 _n-3 ≦ −B, if Δω1 _n-2 changes greatly to the plus side, Δω2 _n > There is a possibility that the condition of A or Δω2 _n ≧ A will be satisfied.
For this reason, in misfire detection based only on the second variation Δω2 _n , misfire may be erroneously diagnosed due to combustion variations.

However, as described above, Δω1 _n-1 > B or Δω1 _n-1 ≧ B and Δω1 _n-3 <−B or Δω1 _n-3 ≦ −B is added as a misfire determination condition. For example, even if Δω2 _n <−A or Δω2 _n ≦ −A and Δω2 _n > A or Δω2 _n ≧ A are satisfied due to combustion variations, the occurrence of misfire is erroneously determined. Can be avoided.
Further, even if the second fluctuation amount Δω2 _n repeatedly changes in the misfire pattern due to combustion variations, if the change pattern of the first fluctuation amount Δω1 _n is not the misfire pattern, the misfire is determined. Therefore, it is possible to distinguish between the single cylinder misfire every time and the fluctuation of the angular velocity in a constant pattern due to combustion variations, and it is possible to detect the misfire every time of the single cylinder with high accuracy.

By the way, in the 4-cylinder engine 101 of the present embodiment, if the # 1 cylinder and the # 2 cylinder, or the # 3 cylinder and the # 4 cylinder misfire each time, the difference “ω _n −ω _n−2 ” “ω _n−1 − One of ω _n-3 ”is obtained as a difference between misfiring cylinders, and the other is obtained as a difference between normal combustion cylinders. The first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n do not fluctuate. Although misfire has occurred, the occurrence of misfire is not determined.

That is, the differences “ω _n −ω _n−2 ” and “ω _n−1 −ω _n-3 ” are obtained as the difference between the cylinders whose cycles (strokes) are shifted from each other by one rotation of the crankshaft 121. For example, if the # 1 cylinder and the # 2 cylinder misfire each time, one of the differences “ω _n −ω _n−2 ” and “ω _n−1 −ω _n-3 ” is the difference between the # 1 cylinder and the # 2 cylinder. The difference in angular velocities results in the difference in angular velocities between the # 3 and # 4 cylinders, and the difference in angular velocities between the misfire cylinder and the normal combustion cylinder cannot be obtained.

Therefore, a third embodiment in which it is possible to detect that the # 1 cylinder and the # 2 cylinder or the # 3 cylinder and the # 4 cylinder are misfired each time will be described with reference to the flowcharts of FIGS.
In the following, a combination of two cylinders whose cycles (strokes) are shifted by one rotation of the crankshaft 121, such as the # 1 cylinder and the # 2 cylinder, or the # 3 cylinder and the # 4 cylinder, will be described. It shall be called a cylinder.

In the flowchart of FIG. 8, in steps S3001 to S3008, processing similar to that in steps S2001 to S2008 of the second embodiment is performed.
However, the steps S3001 to S3008 can be replaced with the steps S1001 to S1006 of the first embodiment. That is, the detection of misfire every time of the # 1 cylinder and # 2 cylinder or the # 3 cylinder and # 4 cylinder of the third embodiment detects the misfire only from the second fluctuation amount Δω2 _n , and the first fluctuation amount The present invention can be applied to both the configuration in which misfire detection is performed from Δω1 _n and the second variation amount Δω2 _n .

If no misfire is detected in the processes of steps S3001 to S3008, the process proceeds to step S3009 for determining whether the front and back cylinders are misfired (multi-cylinder misfire determination means).
Details of the processing in step S3009 are shown in the flowchart of FIG.
First, in step S3101, it is determined whether or not the first variation Δω1 _n is a value near zero (−D <Δω1 _n <D: D << A, B).

If the first fluctuation amount Δω1 _n is a value near zero, the process proceeds to step S3102, and the second fluctuation amount Δω2 _n is a value near zero (−D <Δω1 _n << + α: D << A, B). It is determined whether or not.
The above-mentioned value near zero means that the angular velocity ω _n is stable even if there are fluctuations in the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n due to minute combustion variations between cylinders. It is what you see.

If the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n are values near zero, the process further proceeds to step S3103, where the base air-fuel ratio of the internal combustion engine 101 is biased to the lean side with respect to the target air-fuel ratio. Determine whether or not.
The base air-fuel ratio indicates an air-fuel ratio in a state where feedback control of the fuel injection amount for making the actual air-fuel ratio detected by the air-fuel ratio sensor 122 close to the target is not performed. If there is, the deviation from the feedback correction value to the lean side can be determined, and if the feedback control is in the open control state, it can be determined from the air-fuel ratio detected by the air-fuel ratio sensor 122.

  Further, the leaning of the base air-fuel ratio toward the lean side determines whether or not the deviation of the air-fuel ratio to the lean side, which is expected when two of the four cylinders are misfiring, has occurred. The base air-fuel ratio is biased toward the lean side when the correction value is shifted to the fuel injection amount increase correction side above the predetermined value or when the air-fuel ratio detected by the air-fuel ratio sensor 122 is leaner than the predetermined value. Judge.

If it is determined that the base air-fuel ratio of the internal combustion engine 101 is biased toward the lean side, the process proceeds to step S3104, where the base air-fuel ratio is less than a predetermined value when the fluctuation of the base air-fuel ratio (difference between the previous value and the current value) is less than the predetermined value. Judge whether it is stable to the side.
If it is stable on the lean side, the process proceeds to step S3105, and it is determined whether or not the state where all the conditions in steps S3101 to S3104 are satisfied continues for the determination time or more.

If all the conditions in steps S3101 to S3104 are satisfied for the determination time or longer, the process proceeds to step S3106, and the front and back cylinders (# 1 cylinder and # 2 cylinder, or # 3 Cylinder and # 4 cylinder) are estimated each time.
When the front and back cylinders misfire each time, the first variation Δω1 _n and the second variation Δω2 _n are stabilized near zero, and in the processes in steps S3001 to S3008, misfire is not determined. Because the cylinder continues to misfire, and the air (oxygen) that has passed through the misfired cylinder flows directly into the exhaust system and increases the oxygen concentration in the exhaust, so the exhaust air-fuel ratio is stable and tends to lean Based on the determination in steps S3101 to S3105, it is possible to estimate the occurrence of misfire in the front and back cylinders each time.

However, there is a possibility that the air-fuel ratio is biased to the lean side due to a cause other than misfire, and in the above estimation, the misfire cylinder cannot be specified. Therefore, after step S3107, misfire occurs every time (ignition) Process to determine the front and back cylinders that misfire each time.
In step S3107, as shown in FIG. 10, the fuel injection amount is increased from the previous value for each injection in the order of fuel injection, and the fuel injection amount is increased and changed stepwise (the air-fuel ratio is gradually increased in the rich direction). If the fuel injection amount is increased for all four cylinders, the fuel injection amount is decreased step by step for each injection and returned to the normal value. / Or if the ignition timing is corrected to advance from the previous value for each ignition in accordance with the ignition sequence, and the ignition timing is changed in stages, and ignition is performed at the ignition timing corrected for all four cylinders. Then, the ignition timing is changed stepwise for each ignition to return to the normal value.

Here, the correction of the fuel injection amount and the ignition timing is a correction for changing the generated torque for each cylinder, and by increasing the fuel injection amount and / or the advance angle of the ignition timing step by step, The rotational speed per cylinder (angular velocity ω _n ) in the combustion cylinder of this time slightly changes from that of the cylinder, and conversely, by reducing the fuel injection amount and / or retarding the ignition timing step by step, The rotational speed (angular velocity ω _n ) of each cylinder in the current combustion cylinder is slightly smaller than that of the combustion cylinder.

Since step S3107 is a process for forcibly changing the rotational speed of each cylinder, each cylinder is replaced with the correction of the fuel injection amount / ignition timing or together with the correction of the fuel injection amount and / or the ignition timing. The intake air amount can be changed minutely individually.
As a method of minutely changing the intake air amount of each cylinder individually, a throttle valve is provided for each cylinder and the opening degree is corrected for each cylinder, or the opening characteristics of the intake valve (valve lift amount, valve operating angle, There is a method of changing the opening / closing timing.

In step S3108, the front and back cylinders that misfire each time are identified based on the difference Δω _n = ω _n −ω _n−2 that is sequentially calculated during the correction process.
For example, when both the # 1 cylinder and the # 2 cylinder misfire each time, if the correction process in step S3107 is performed, the difference Δω _n changes as shown in FIG.
Since the example shown in FIG. 10 shows a case where the # 1 cylinder and the # 2 cylinder are misfired every time, even if the fuel injection amount and the ignition timing are first corrected for the # 1 cylinder, the correction result is The cylinder-specific rotation speed (angular speed ω _n ) of the # 1 cylinder is not increased or decreased, and there is no difference with respect to the cylinder-specific rotation speed (angular speed ω _n ) of the # 2 cylinder ignited before the correction is started. The difference Δω _n is calculated to be substantially zero.

On the other hand, since the next # 3 cylinder in the ignition order is a cylinder that is normally ignited and burned, if the fuel injection amount and / or the ignition timing are corrected, the rotational speed (angular speed) of the # 4 cylinder before the correction start Since the rotational speed per cylinder (angular speed ω _n ) slightly increases in comparison with ω _n ), the difference Δω _n increases from the previous substantially zero state to the plus side.
Further, since the next cylinder # 2 is a misfire cylinder every time, even if the fuel injection amount and the ignition timing are corrected, the rotational speed (angular speed ω _n ) for each cylinder does not change, and the difference Δω _n is compared. Since the cylinder specific rotation speed (angular velocity ω _n ) of the # 1 cylinder after the start of correction is also a value that does not show a change corresponding to the correction, the difference Δω _n is calculated to be substantially zero.

Further, the next # 4 cylinder is a cylinder that is normally ignited and burned, and moreover, the fuel injection amount increase and / or the ignition timing advance angle is larger than in the previous # 3 cylinder. Therefore, the cylinder-specific rotational speed (angular speed ω _n ) of the # 4 cylinder is larger than the cylinder-specific rotational speed (angular speed ω _n ) of the # 3 cylinder to be compared in the difference Δω _n , and the difference Δω _n is increased and changed from the previous substantially zero state to the plus side.

Further, from the next cylinder # 1, the fuel injection amount decrease correction and / or the ignition timing retard correction is started, but since the # 1 cylinder is a misfire cylinder every time, the fuel injection amount increase correction and / or Alternatively, the difference Δω _n is calculated to be substantially zero as in the case where the ignition timing advance angle correction is performed.
On the other hand, for the # 3 cylinder after the start of the reduction / retarding correction, the amount of increase in the fuel injection amount and / or the advance of the ignition timing is reduced compared to the case of the # 4 cylinder before the previous time. The rotational speed for each cylinder (angular speed ω _n ) is smaller than the rotational speed for each cylinder (angular speed ω _n ) of # 4 cylinder, and the difference Δω _n decreases from the previous substantially zero state to the minus side. Become.

Since the # 2 cylinder after the start of the reduction / retarding correction is a misfiring cylinder every time, the difference Δω _n is the same as when the fuel injection amount increase correction and / or the ignition timing advance correction is performed. It is calculated to be substantially zero.
In the next cylinder # 4, the fuel injection amount reduction correction and / or the ignition timing retardation correction are advanced step by step with the same step width as that during the increase correction and / or advance correction. In addition, the ignition timing returns to the normal value.

Here, the cylinder-specific rotational speed (angular speed ω _n ) of the # 3 cylinder to be compared in the difference Δω _n is a value in a state in which the increase amount and / or the advance angle correction amount remain, so that the fuel injection the amount and than cylinder rotational speed of the fourth cylinder in a state where the ignition timing is returned to a normal value (angular velocity omega _n), cylinder rotational speed of the cylinder # 3 (the angular velocity omega _n) is high, the difference [Delta] [omega _n is It will decrease from the previous substantially zero state to the minus side.

As described above, even when the fuel injection amount and / or the ignition timing are corrected, the difference Δω _n obtained between the front and back cylinders misfiring each time is kept substantially zero.
On the other hand, the difference Δω _n obtained between the normally burning front and back cylinders is increased and / or advanced while the fuel injection amount is increased and / or the ignition timing is advanced. Indicates a positive value corresponding to the step width (a slight increase change in the rotational speed (angular velocity ω _n ) for each cylinder corresponding to the step width), and the fuel injection amount reduction correction and / or the ignition timing retardation correction are performed. In the meantime, a negative value corresponding to the step width of the reduction correction and / or the retard correction (a slight decrease change in the rotational speed (angular velocity ω _n ) corresponding to the step width) is shown.

Therefore, by comparing the above characteristics with the actual change in the difference Δω _n , it is possible to identify the front and back cylinders that misfire each time. As shown in FIG. 10, correction is started from the # 1 cylinder. Specific correction conditions in the case where the correction is made to increase the angular velocity for all cylinders and then the correction is made to decrease the angular velocity for all cylinders to restore the original state to the # 4 cylinder will be shown below.

When all of the following conditions (a) to (h) are satisfied, misfire is determined every time for the # 1 cylinder and the # 2 cylinder. However, the determination values C and D shown in the following conditions are C >> D.
(A): “Δω _n <(≦) −C”
(B): “−D <(≦) Δω _n−1 <(≦) D”
(C): “Δω _n-2 <(≦) −C”
(D): “−D <(≦) Δω _n−3 <(≦) D”
(E): “Δω _n-4 > (≧) C”
(F): “−D <(≦) Δω _n−5 <(≦) D”
(G): “Δω _n-6 > (≧) C”
(H): “−D <(≦) Δω _n−7 <(≦) D”
Of the above determination condition, that the difference [Delta] [omega _n is determined whether the value of the region sandwiched between the judged value D and the determination value -D is the difference [Delta] [omega _n is equal to or substantially zero Further, the determination of whether the value is a negative value lower than the determination value −C or a positive value higher than the determination value C is determined by the angular velocity ω _n corresponding to the correction of the fuel injection amount and / or the ignition timing. It is determined whether or not a change has occurred.

Therefore, the absolute value of the judgment value D is a minute value, and the magnitude of the judgment value C is set based on the change rate of the angular velocity commensurate with the correction of the fuel injection amount and / or the ignition timing.
Further, when all of the following conditions (i) to (p) are satisfied, misfire is determined every time for the # 3 cylinder and the # 4 cylinder.
(I): “−D <(≦) Δω _n <(≦) D”
(J): “Δω _n-1 <(≦) −C”
(K): “−D <(≦) Δω _n−2 <(≦) D”
(L): “Δω _n-3 <(≦) −C”
(M): “−D <(≦) Δω _n−4 <(≦) D”
(N): “Δω _n-5 > (≧) C”
(O): “−D <(≦) Δω _n-6 <(≦) D”
(P): “Δω _n-7 > (≧) C”
According to the third embodiment, it is possible to determine misfiring of the front and back cylinders that cannot be determined on the basis of the first variation amount Δω1 _n and the second variation amount Δω2 _n , thereby improving the reliability and detection accuracy of misfire detection. Can be made.

Further, in the third embodiment, misfire is determined each time in the front and back cylinders based on a minute change in the rotational speed (angular speed ω _n ) for each cylinder accompanying the correction of the fuel injection amount and the ignition timing. In the low speed range, the front and back cylinders can be identified without causing the driver to feel uncomfortable.
When all of the above conditions (a) to (h) are satisfied, the misfire is determined every time for the # 1 cylinder and the # 2 cylinder, and when there is a condition that is not satisfied among the above conditions (a) to (h). , # 3 cylinder and # 4 cylinder can be determined each time misfire, conversely, when all of the above conditions (i) to (p) is satisfied, # 3 cylinder and # 4 cylinder each time misfire is determined, When there is a condition that does not hold among the above conditions (i) to (p), it is possible to determine misfire of the # 1 cylinder and the # 2 cylinder every time.

In addition, it is not necessary to start correction of the fuel injection amount and the ignition timing from the # 1 cylinder, and it is not necessary to increase or decrease the correction degree sequentially in accordance with the ignition order. Further, the fuel injection amount and the ignition timing are corrected. It is possible to specify the front and back cylinders misfiring each time from the first variation amount Δω1 _n and / or the second variation amount Δω2 _n .
That is, the front and back cylinders misfiring each time can be specified by whether or not the change in the rotational speed (angular speed) for each cylinder in accordance with the correction of the fuel injection amount and the ignition timing is indicated. The method can be set as appropriate.

Next, a fourth embodiment for identifying the front and back cylinders misfiring each time by another method will be described with reference to the flowchart of FIG.
In the flowchart of FIG. 11, in steps S4101 to S4106, each time a misfire of the front and back cylinders (# 1 cylinder and # 2 cylinder, or # 3 cylinder and # 4 cylinder) is estimated as in steps S3101 to S3106. Judgment is made as to whether or not it is in a state to be performed.

Then, if misfire is estimated each time in the front and back cylinders in step S4106, the process proceeds to step S4107, where a process of stopping only one injection of the specific cylinder is performed, and then the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 calculated thereafter are performed. Remember _n .
Then, in step S4107, on the basis of the first variation amount .DELTA..omega.1 _n and the second variation amount ?? 2 _n, which is calculated after the fuel cut, it identifies the front and back a misfired cylinder and a each time.

For example, in the case where the # 3 cylinder and the # 4 cylinder are misfired every time, when the fuel injection of the # 1 cylinder is stopped only by one injection in such a state (the fuel cut is performed for one injection for the # 1 cylinder) Then, as for the # 1 cylinder, the rotational speed (angular speed ω _n ) for each cylinder drops as in the case of misfire (see FIG. 12).
Here, since the normal combustion state continues for the # 2 cylinder whose cycle is shifted by one crankshaft rotation with respect to the # 1 cylinder, the rotation speed for each cylinder between the # 1 and # 2 cylinders. A difference occurs in (angular velocity ω _n ), and as a result, fluctuation occurs in the first fluctuation amount Δω 1 _n and the second fluctuation amount Δω 2 _n .

The fluctuations in the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n occur similarly when a fuel cut is executed for the # 2 cylinder instead of the # 1 cylinder.
On the other hand, for one of the front and back cylinders that misfires each time, even if the fuel injection is stopped only by one injection, it is the cylinder that originally misfired. Therefore, the fuel injection is stopped at the rotational speed for each cylinder (angular velocity ω _n ). There is no effect, and the first variation amount Δω1 _n and the second variation amount Δω2 _n do not vary.

Therefore, simply, it is determined whether or not the cylinder-by-cylinder rotation speed (angular velocity ω _n ) of one cylinder that has performed fuel cut has dropped, and if a drop in rotation occurs, the front and back cylinders that do not include the cylinder Therefore, it can be determined that the front and back cylinders including the cylinder are misfired each time if the rotation is not reduced.
Further, in order to discriminate front and back cylinders misfiring each time with higher accuracy from the fluctuations of the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n caused by the fuel cut, the following conditions are determined.

First, only one injection is stopped for the # 1 cylinder (or # 2 cylinder), and after the four ignitions (after sampling time-series four data), whether or not the following condition (X) is satisfied: Judging.
(X): “Δω2 _n <(≦) −A, Δω1 _n−1 > (≧ B), Δω2 _n > (≧) A, and Δω1 _n-3 <(≦) −B”
Note that the condition (X) is the same as the determination in steps S2004 to S2007.

As a result of stopping only one injection for the # 1 cylinder (or # 2 cylinder), the fluctuation of the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n shows the behavior at the time of single cylinder misfire, and the condition (X) is satisfied If it is determined that the misfire may be specified every time for the # 3 and # 4 cylinders at that time, it is further determined that the condition (X) is satisfied in order to ensure the specification of the misfiring cylinder. After a sufficient time has passed for the rotational fluctuations to converge from this point, this time, only one injection is stopped for the # 3 cylinder (or # 4 cylinder).

Then, after stopping only one injection for the # 3 cylinder (or # 4 cylinder), it is determined whether or not the following condition (Y) is satisfied after four ignitions.
(Y): “−D <(≦) Δω1 _n <(≦) D, −D <(≦) Δω _n−1 <(≦) D, and −D <(≦) Δω _n−2 < (≦) D and −D <(≦) Δω _n−3 <(≦) D ”
Here, the condition (Y) is used to determine whether or not the first variation Δω1 _n continues and is stable near zero. The front and back cylinders that misfire each time are the # 3 cylinder and the # 4 cylinder. In the case of a cylinder, the fuel cut does not affect the angular velocity ω _n , and the first variation amount Δω 1 _n and the second variation amount Δω 2 _n hold values close to zero. This is consistent with the establishment of the condition (X), and from the establishment of both the conditions (X) and (Y), it is determined that the front and back cylinders misfiring each time are the # 3 cylinder and the # 4 cylinder. Can do.

However, for the # 1 or # 2 cylinders, the front and back cylinders of the # 3 and # 4 cylinders can be obtained only by satisfying the above condition (Y) without determining the above condition (X) based on the fuel cut. Can be determined to have misfired each time.
In addition, it is first determined whether or not the condition (Y) is satisfied, and then it can be determined whether or not the condition (X) is satisfied.

On the other hand, whether or not the misfiring front and back cylinders are the # 1 cylinder and the # 2 cylinder is determined as follows.
First, for the # 1 cylinder (or # 2 cylinder), only one injection is stopped, and after the four ignitions, it is determined whether or not the following condition (α) is satisfied.
(Α): “−D <(≦) Δω1 _n <(≦) D, −D <(≦) Δω _n−1 <(≦) D, and −D <(≦) Δω _n−2 < (≦) D and −D <(≦) Δω _n−3 <(≦) D ”
If the # 1 and # 2 cylinders are front and back cylinders that misfire each time, the fuel-cutting speed (angular velocity ω _n ) is not affected by the fuel cut, and the first fluctuation amount Δω 1 _n continues to be zero. If the condition (α) is satisfied, it can be determined that the # 1 and # 2 cylinders are front and back cylinders that misfire each time.

Next, in order to support the determination that the # 1 and # 2 cylinders are the front and back cylinders misfiring each time, it is sufficient for the rotational fluctuations to converge after the above condition (α) is determined. After waiting for the time, this time, for the # 3 cylinder (or # 4 cylinder), only one injection is stopped, and it is determined whether or not the following condition (β) is satisfied after the fourth ignition.
(Β): “Δω2 _n <(≦) −A, Δω1 _n−1 > (≧ B), Δω2 _n > (≧) A, and Δω1 _n-3 <(≦) −B”
As a result of stopping only one injection for the # 3 cylinder (or # 4 cylinder), the fluctuation of the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n shows the behavior at the time of single cylinder misfire, and the condition (β) is satisfied. Then, it is determined that the # 1 cylinder and the # 2 cylinder are front and back cylinders misfiring each time.

Again, based on the establishment of one of the conditions (α) and (β), the front and back cylinders misfiring each time can be determined to be the # 1 cylinder and the # 2 cylinder, and the conditions (α), The determination order of (β) can be reversed.
As described above, if the fuel cut is performed and the misfiring front and back cylinders are determined each time based on whether or not a change in the angular velocity ω _n corresponding to the fuel cut occurs, as in the third embodiment, the fuel Compared to correcting the injection amount and ignition timing, the angular velocity change in a normally ignited cylinder is more clearly generated, so misfires occur every time even at high revolutions, where rotation fluctuations are unlikely to occur. The front and back cylinders can be identified with high accuracy.

Further, during high revolution, the front and back cylinders are misfired every time, and even if fuel cut is performed, it is possible to avoid the internal combustion engine 101 from being stopped. Therefore, the misfire front and back based on the fuel cut according to the fourth embodiment. The cylinder specification can be executed only in the high rotation region of the internal combustion engine 101.
Accordingly, when the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n hold values close to zero, the air-fuel ratio is stable on the lean side, and the occurrence of misfiring in each of the front and back cylinders is estimated. If the engine rotation speed is in the low rotation range, the front and back cylinders misfiring each time are specified by the processing shown in the flowchart of FIG. 9, and if the engine rotation speed is in the high rotation range, the processing shown in the flowchart of FIG. The front and back cylinders misfiring each time can be specified.

In addition, the estimation of misfire for each of the front and back cylinders based on the fact that the first fluctuation amount Δω1 _n and the second fluctuation amount Δω2 _n hold values close to zero and the air-fuel ratio is stable on the lean side is omitted. Whether or not the front and back cylinders misfire each time by executing the processing shown in the flowchart of FIG. 9 or 11 in a state where no misfire is detected based on the first variation Δω1 _n and the second variation Δω2 _n . Can be judged.

In each of the above embodiments, when the occurrence of misfire is detected, if misfire occurs every time, the fuel injection to the cylinder is stopped, and if misfire occurs at a frequency exceeding the allowable level, the driver It is preferable to warn the user, limit the operation region of the internal combustion engine 101 to the low rotation / low load side, or store the misfire occurrence as a failure history.
In the present embodiment, the internal combustion engine 101 is a four-cylinder engine. However, the misfire detection apparatus according to the present invention can be applied to a six-cylinder engine, an eight-cylinder engine, and the like. in a six-cylinder engine phase difference is 120 deg in crank angle determines the cylinder rotation speed (angular velocity omega _n) for each 120 deg, the first variation amount _{Δω1 n, Δω1 n = (ω} n -ω n-3 ) − (Ω _n−1 −ω _n-4 ).

In the calculation formula of the first variation Δω1 _n , “ω _n −ω _n-3 ” and “ω _n−1 −ω _n-4 ” are both out of cycle by one rotation of the crankshaft. The difference in the rotational speed of each cylinder (angular speed ω _n ) between the cylinders is obtained, and as in the case of the four cylinders, the influence of mechanical variation of the crank angle sensor 117 can be eliminated and the misfire cylinder can be detected. it can.
Further, in an 8-cylinder engine in which the stroke phase difference between cylinders is 90 deg in crank angle, the rotation speed (angular velocity ω _n ) for each cylinder is obtained every 90 deg, and the first variation Δω1 _n is calculated as Δω1 _n = (ω _{n −} ω _n−4 ) − (ω _n−1 −ω _n−5 ).

Also in this case, both “ω _n −ω _n-4 ” and “ω _n−1 −ω _n-5 ” are the rotation speeds for each cylinder between cylinders whose cycles are shifted by one rotation of the crankshaft ( The difference in the angular velocity ω _n ) is obtained, and as in the case of the four cylinders, the influence of the mechanical variation of the crank angle sensor 117 can be eliminated and the misfired cylinder can be detected.
Further, instead of calculating the difference between the rotational speeds per cylinder (angular speed ω _n ) as a difference between the values detected by being shifted by one revolution of the crankshaft, an integer n (≧≧ 2) It can be calculated as the difference between the values detected by shifting by a factor of two.

1 is a system diagram of an internal combustion engine in an embodiment of the present invention. The flowchart which shows 1st Embodiment of the misfire detection which concerns on this invention. The figure which shows the detection area | region of the angular velocity in embodiment of this invention. The flowchart which shows the detection process of the angular velocity in embodiment of this invention. The flowchart which shows the detection process of the angular velocity in embodiment of this invention. Angular velocity omega _n, the first variation amount .DELTA..omega.1 _n and the second variation amount ?? 2 _n time chart showing change in time of a misfire in the embodiment of the present invention. The flowchart which shows 2nd Embodiment of the misfire detection which concerns on this invention. The flowchart which shows 3rd Embodiment of the misfire detection which concerns on this invention. The flowchart which shows the detail of the misfire detection of the front and back cylinder in the said 3rd Embodiment. The time chart which shows the correlation with the change of the correction state of the fuel injection amount and ignition timing in the said 3rd Embodiment, and the change of difference (DELTA) omega _n . The flowchart which shows 4th Embodiment of the misfire detection which concerns on this invention. The fourth timing chart showing a variation of the first variation amount .DELTA..omega.1 _n and the second variation amount ?? 2 _n accompanying fuel cut in the embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 101 ... Internal combustion engine, 103 ... Electronically controlled throttle device, 112 ... Fuel injection valve, 113 ... Engine control unit (ECU), 114 ... Spark plug, 115 ... Ignition coil, 117 ... Crank angle sensor, 121 ... Crankshaft (output shaft) )

Claims (5)

Cylinder-specific rotation speed detecting means for detecting the cylinder-specific rotation speed of the output shaft of the multi-cylinder internal combustion engine;
Detecting means for detecting a change in the rotational speed of each cylinder between cylinders whose cycle is shifted by an integral multiple of one rotation of the output shaft;
First fluctuation amount calculating means for calculating the fluctuation amount of the change as a first fluctuation amount;
Second fluctuation amount calculating means for calculating the fluctuation amount of the first fluctuation amount as a second fluctuation amount;
Misfire determination means for determining the presence or absence of misfire based on the second variation amount;
A misfire detection apparatus for a multi-cylinder internal combustion engine, comprising: Cylinder-specific rotation speed detecting means for detecting the cylinder-specific rotation speed of the output shaft of the multi-cylinder internal combustion engine;
Detecting means for detecting a change in the rotational speed of each cylinder between cylinders whose cycle is shifted by an integral multiple of one rotation of the output shaft;
First fluctuation amount calculating means for calculating the fluctuation amount of the change as a first fluctuation amount;
Second fluctuation amount calculating means for calculating the fluctuation amount of the first fluctuation amount as a second fluctuation amount;
Misfire determination means for determining the presence or absence of misfire based on the first variation amount and the second variation amount;
A misfire detection apparatus for a multi-cylinder internal combustion engine, comprising: Cylinder-specific rotation speed detecting means for detecting the cylinder-specific rotation speed of the output shaft of the multi-cylinder internal combustion engine;
Detecting means for detecting a change in the rotational speed of each cylinder between cylinders whose cycle is shifted by an integral multiple of one rotation of the output shaft;
First fluctuation amount calculating means for calculating the fluctuation amount of the change as a first fluctuation amount;
Second fluctuation amount calculating means for calculating the fluctuation amount of the first fluctuation amount as a second fluctuation amount;
Misfire determination means for determining the presence or absence of misfire based on the second variation amount;
A plurality of cylinder misfire determination means for determining misfire each time in a plurality of cylinders when the misfire is not determined by the misfire determination means and the air-fuel ratio of the multi-cylinder internal combustion engine is biased to the lean side;
A misfire detection apparatus for a multi-cylinder internal combustion engine, comprising:   The multi-cylinder misfire determination means corrects a fuel injection amount and / or ignition timing for each cylinder, and identifies a plurality of misfire cylinders based on fluctuations in the rotation speed for each cylinder generated as a result of the correction. The misfire detection apparatus for a multi-cylinder internal combustion engine according to claim 3.   The multi-cylinder misfire determination means performs a fuel cut temporarily in a part of the cylinders, and identifies a plurality of misfire cylinders on the basis of fluctuations in the rotation speed for each cylinder generated in association with the fuel cut. The misfire detection apparatus for a multi-cylinder internal combustion engine according to claim 3.

Images