JPH04265475A - Mis-fire judgment for respective cylinders of engine - Google Patents

Abstract

PURPOSE:To detect mis-fire judgment for respective cylinders accurately. CONSTITUTION:A mis-fire judgment level DELTANLEVEL which is enclosed in a mis-fire judgment level map MP DELTANLEVEL is updated with weighted mean which is preset based on a combustion condition judgment value DELTAN#i-1 detected at the time of combustion cut, and the updated mis-fire judgment level NLEVEL is compared with a combustion condition judgment value DELTAN#i-1 which is obtained at the time of fuel injection to judge mis-fire. Because the mis-fire judgment level DELTANLEVEL is corrected based on the combustion condition judgment value DELTAN#i-1 which is obtained at the time of fuel cut, friction elements which are different from respective cylinders each other are eliminated. It is thus possible to improve mis-fire detection accuracy.

Description

[Detailed description of the invention]

0001

[Industrial Field of Application] The present invention is an engine that sets a misfire judgment level for each cylinder based on rotational fluctuations during fuel cut, and individually judges the misfire status of each cylinder based on this misfire judgment level. This invention relates to a method for determining misfire by cylinder.

0002

[Prior art and problems to be solved by the invention] Generally,
Ideally, combustion in a multi-cylinder engine should be carried out through the same process every cycle in order to obtain stable output, but in a multi-cylinder engine, there are Uneven distribution of intake air,
Combustion tends to vary due to synergistic effects such as: (slight differences in combustion temperature between cylinders caused by the cooling route); (2) manufacturing variations in combustion chamber volume, piston shape, etc. of each cylinder;

Conventionally, combustion fluctuations between cylinders have been suppressed to a minimum by cylinder-specific air-fuel ratio control and ignition timing control. If the injector, spark plug, etc. deteriorate or malfunction, it is likely to cause intermittent misfires and a decrease in output.

In a multi-cylinder engine, even if intermittent misfire occurs in one cylinder, it is often operated without noticing, and it is also difficult to determine whether the cause of the misfire is a temporary one or whether the injector It is difficult to determine while driving whether the problem is caused by deterioration of the spark plug or the like.

[0005] Therefore, for example, Japanese Patent Laid-Open No. 61-2589
In Publication No. 55, the difference between the minimum value and the maximum value of the engine rotation speed of the cylinder before one combustion stroke is compared with the difference between the minimum value and the maximum value of the engine rotation speed of the cylinder in the current combustion stroke, and this comparison value is The combustion state of the cylinder in the current combustion stroke is determined based on whether it falls within a preset reference value, and if combustion abnormality occurs more than a predetermined number of times, it is determined that a misfire has occurred and a warning is issued.

However, in this prior art, combustion fluctuations in each cylinder are determined from the difference between the minimum engine rotation speed and the maximum engine rotation speed of the combustion stroke cylinder, but the engine rotation speed during combustion increases rapidly. Furthermore, since a relatively large load is placed on the engine, the fluctuations in acceleration become large, making it difficult to determine the maximum value of the engine speed and increasing the accuracy error when determining a misfire. .

Incidentally, the combustion characteristics of an engine vary not only between cylinders but also from engine to engine due to manufacturing errors in parts.

If the reference value for comparing rotational fluctuations is set as an absolute value as in the prior art described above, the reference value will vary relative to each engine due to variations in the combustion characteristics of each engine. In this case, it becomes difficult to accurately grasp combustion abnormalities.

[0009] In engines with a small number of cylinders, the combustion interval between cylinders is relatively long, so the difference in rotational speed is large, and even if the reference value is set as an absolute value, variations in combustion characteristics from engine to engine have a large effect on misfire determination. Although it does not affect
In an engine with a large number of cylinders, the combustion interval between cylinders becomes shorter, and the difference in rotational fluctuation becomes smaller. Therefore, if the determination level (reference value) is set in advance as an absolute value, the misfire determination accuracy will be affected by variations in the combustion characteristics of each engine. will have a major impact on

[0010] Particularly in a high rotation range, the variation difference becomes small, so if the determination level varies relative to each engine, it becomes extremely difficult to make an accurate misfire determination.

For example, in Japanese Patent Application Laid-Open No. 59-82534, the instantaneous engine speed NL#i before the combustion stroke of each cylinder #i (i=1 to 4 in the case of 4 cylinders) and the instantaneous engine speed NL#i after the combustion stroke are disclosed. The differential rotation ΔN#i (ΔN#i=NH#i - NL#i), which is the difference from the instantaneous engine rotational speed NH#i, is determined for each cylinder, and then the differential rotation ΔN of each cylinder #i is determined for each cylinder.
The combustion state of each cylinder #i is grasped by comparing the all-cylinder average value ΔNA of #i with the differential rotation ΔN#i of each cylinder #i.

However, in this prior art, since the reference value is the all-cylinder average differential rotation ΔNA, this all-cylinder average differential rotation ΔNA itself is always likely to fluctuate depending on the combustion condition; Then, the combustion state of each cylinder estimated based on this all-cylinder average differential rotation ΔNA includes the combustion condition factors of other cylinders when setting the above-mentioned all-cylinder average differential rotation ΔNA. It becomes difficult to accurately identify misfires.

[0013] Furthermore, as disclosed in Japanese Patent Laid-Open No. 63-268956, a misfire is determined based on the difference in momentum before and after combustion (in the publication, difference in rotational circumference). , it is simply judged as a misfire, but for example, when driving at a constant speed, the difference in momentum may become negative due to the influence of engine friction even though there is no misfire (normal combustion), and there is a risk of misjudgment. There is.

The present invention has been made in view of the above circumstances, and eliminates the misfire condition from including the combustion state factors of other cylinders, and eliminates not only combustion variations between cylinders, but also manufacturing variations from engine to engine, and The object of the present invention is to provide a method for determining engine misfires by cylinder, which can accurately detect engine misfires without being affected by engine friction.

[0015]

[Means for Solving the Problems] (1) In order to achieve the above object, the first engine cylinder-by-cylinder misfire determination method according to the present invention is based on the momentum and current of a predetermined section of a cylinder before two combustion strokes at the time of fuel cut. The combustion state discrimination value of the cylinder before one combustion stroke is calculated from the difference between the average value of the momentum of the cylinder in the predetermined section of the combustion stroke and the momentum of the cylinder in the predetermined section of the cylinder before one combustion stroke, and the above momentum of the cylinder before one combustion stroke is calculated. 1 read out as a parameter
The misfire determination level stored in a predetermined address of the misfire determination level map for the cylinder before the combustion stroke is updated with the value corrected by the combustion state determination value, and then the momentum of the predetermined section of the cylinder before the two combustion strokes during fuel injection is calculated. The combustion state discrimination value of the cylinder one combustion stroke before is calculated from the difference between the average value of the momentum of the cylinder in the current combustion stroke in a predetermined section and the momentum of the cylinder one combustion stroke before. The state determination value is compared with the misfire determination level stored in a predetermined address of the misfire determination level map for the cylinder one combustion stroke ago, which is read out using the momentum of the cylinder one combustion stroke ago as a parameter. It is characterized by determining a misfire state.

(2) In order to achieve the above object, the second engine misfire determination method according to the present invention is based on the momentum of a predetermined section of the cylinder in the current combustion stroke and the predetermined section of the cylinder one combustion stroke before the fuel cut. The combustion state determination value of the current combustion stroke cylinder is calculated from the difference with the momentum of the current combustion stroke cylinder, and the misfire determination value stored at a predetermined address of the misfire determination level map of the current combustion stroke cylinder is read out using the momentum of the current combustion stroke cylinder as a parameter. The level is updated with a value corrected with the above combustion state discrimination value, and then the combustion of the current combustion stroke cylinder is determined based on the difference between the momentum in a predetermined section of the current combustion stroke cylinder during fuel injection and the momentum in a predetermined section of the cylinder one combustion stroke before. The state determination value is determined, and the combustion state determination value of the current combustion stroke cylinder is used as a parameter, and the misfire determination level stored at a predetermined address of the misfire determination level map of the current combustion stroke cylinder is read out using the above-mentioned momentum of the current combustion stroke cylinder as a parameter. This feature is characterized in that the misfire state of the cylinder in the current combustion stroke is determined by comparing the two.

[0017]

[Function] (1) In the above-mentioned first engine cylinder-by-cylinder misfire determination method, when determining a misfire in a cylinder one combustion stroke before, first, the friction factor exerted on the relevant cylinder of the engine is detected. The average value of momentum in a predetermined section of the combustion stroke cylinders before and after the cylinder before one combustion stroke, and the above-mentioned 1
A combustion state determination value is obtained from the difference between the momentum of the cylinder in a predetermined section before the combustion stroke.

[0018] Then, the misfire judgment level stored in the address of the misfire judgment level map for each cylinder whose parameter is the momentum of the cylinder one combustion stroke before is corrected by the combustion state judgment value, and with this correction value, the above-mentioned The misfire determination level stored at the address in the misfire determination level map is updated to learn this misfire determination level.

After that, a combustion state determination value is determined from the difference between the average value of the momentum in a predetermined section of the combustion stroke cylinders before and after the cylinder before and after one combustion stroke during fuel injection, and the momentum in the predetermined section of the cylinder before one combustion stroke. The misfire determination level stored at a predetermined address in the misfire determination level map for the cylinder one combustion stroke before is read out using the combustion state discrimination value of the cylinder one combustion stroke prior and the momentum of the cylinder one combustion stroke prior as a parameter. If the combustion state determination value is smaller than the misfire determination level, it is determined that the cylinder has misfired. (
2) In the second engine misfire determination method for each cylinder,
When determining a misfire in a combustion stroke cylinder, first, in order to detect the friction factor that affects the cylinder in the engine,
A combustion state determination value is obtained from the difference in momentum in a predetermined section of the combustion stroke cylinder at the time of fuel cut and the cylinder one combustion stroke before.

[0020] Then, the misfire judgment level stored in the misfire judgment level map for each cylinder at an address whose parameter is the momentum of the relevant combustion stroke cylinder is corrected by the above-mentioned combustion state discrimination value, and the above-mentioned misfire judgment is performed using this correction value. The misfire determination level stored at the address in the level map is updated to learn this misfire determination level.

[0021] Thereafter, a combustion state discrimination value is obtained from the difference in momentum in a predetermined section between the cylinder in the combustion stroke and the cylinder one combustion stroke before the fuel injection, and the combustion state discrimination value for the cylinder in the combustion stroke in question and the cylinder in the combustion stroke in question are determined. The misfire determination level stored in a predetermined address of the misfire determination level map of the relevant combustion stroke cylinder is read out as a parameter, and if the combustion state determination value is smaller than the misfire determination level, the relevant It is determined that the cylinder has misfired.

[0022]

Embodiments Hereinafter, embodiments of the present invention will be explained based on the drawings.

1 to 10 show a first embodiment of the present invention, FIGS. 1 to 3 are flowcharts showing a cylinder-by-cylinder misfire discrimination procedure, FIG. 4 is a flowchart showing a fuel cut discrimination procedure,
FIG. 5 is a schematic diagram of the engine control system, FIG. 6 is a front view of the crank rotor and crank angle sensor, FIG. 7 is a front view of the cam rotor and cam angle sensor, and FIG. 8 is a diagram showing in-cylinder pressure fluctuations, crank pulses, cam pulses, and A time chart showing engine rotation fluctuations, FIG. 9 is a conceptual diagram of a misfire determination level map,
FIG. 10 is a conceptual diagram showing a basic misfire determination method for each cylinder.

(Configuration) Reference numeral 1 in FIG. 5 is an engine;
In the figure, a four-cylinder horizontally opposed engine is shown. An intake manifold 3 is communicated with an intake port 2a formed in a cylinder head 2 of this engine 1, and a throttle chamber 5 is communicated upstream of this intake manifold 3 via an air chamber 4. An air cleaner 7 is attached via a pipe 6.

Furthermore, the air cleaner 7 of the intake pipe 6
An intake air amount sensor (a hot wire air flow meter in the figure) 8 is interposed immediately downstream of the throttle valve 5 provided in the throttle chamber 5.
A throttle opening sensor 9a and an idle switch 9b for detecting fully closed throttle valve are connected to the throttle opening sensor 9a. Further, a multi-point injector (hereinafter abbreviated as "MPI") 10 is disposed immediately upstream of each intake port 2a of each cylinder of the intake manifold 3. Further, reference numeral 11 is a fuel tank, and 12 is a fuel pump that supplies fuel to the MPI 10.

A crank rotor 15 is pivotally attached to the crankshaft 13 of the engine 1, and an electromagnetic pickup or the like is mounted on the outer periphery of the crank rotor 15 for detecting a protrusion (which may be a slit) corresponding to a predetermined crank angle. A crank angle sensor 16 is provided oppositely, and further,
A cam rotor 17 is connected to a camshaft 14 that rotates by 1/2 relative to the crankshaft 13, and a cam angle sensor 18 is provided on the outer periphery of the cam rotor 17.

As shown in FIG. 6, the crank rotor 1
5 has protrusions (slits may also be used) 15a, 15b,
15c is formed. These protrusions 15a, 15b,
15c is the compression top dead center (BTDC) of each cylinder θ1, θ
2, θ3, and the protrusions 15a, 1
From the transit time between 5b and 5b, the period f1,2 (here,
f = 1/ω ω: angular velocity), and also the protrusion 15
The period f2,3 is calculated from the transit time between b and 15c. Further, the protrusion 15b indicates a reference crank angle when setting the ignition timing.

Generally, the ignition timing during idling is BT.
The temperature is around 20°C DC, and even if ignition occurs at this crank angle, the combustion pressure will not rise sharply until about 10°C thereafter.

Further, as shown in FIG. 8, in the embodiment, the opening timing of the exhaust valve of each cylinder is set to be slightly retarded than the ignition reference crank angle BTDCθ2 of the next combustion cylinder. Generally, the combustion pressure immediately after the exhaust valve is opened drops rapidly, so the combustion pressure has almost no effect at the crank angle BTDCθ3.

Therefore, if the crank angle θ3 of the protrusion 15c is set to be more advanced than BTDC10°C, the crank angles BTDCθ2, θ of the protrusions 15b and 15c will be
The section between 3 and 3 is almost unaffected by the combustion between the cylinders, that is, the section in which no work is done by combustion in the combustion stroke cylinder.

Further, as shown in FIG. 7, on the outer periphery of the cam rotor 17, a protrusion (slit may be used) 1 is provided for cylinder discrimination.
7a, 17b, and 17c are formed. A protrusion 17a is formed at a position θ4 after the compression top dead center (ATDC) of the #3 and #4 cylinders, and a protrusion 17b is formed of three protrusions, the first of which is formed at a position after the compression top dead center (ATDC) of the #1 cylinder. After the match (ATD)
C) The protrusion 17c is formed at the position θ5, and the protrusion 17c is formed at the position θ5.
The first protrusion is formed at a position θ6 after the compression top dead center (ATDC) of the #2 cylinder.

[0032] In the example shown in the figure, θ1 = 97°C A,
θ2 = 65℃A, θ3 = 10℃A, θ4 = 20℃
A, θ5 = 5℃A, θ6 = 20℃A, θ(2-3)
=55°C A, and with this arrangement, as shown in FIG. 8, for example, the cam angle sensor 18
When the cam pulse b) is detected, it can be determined that the crank pulse subsequently detected by the crank angle sensor 16 is a signal indicating the crank angle of the #3 cylinder. In addition, when the cam pulse of θ4 (protrusion 17a) is detected after the cam pulse of θ5, the subsequent crank angle sensor 16
It can be determined that the detected crank pulse indicates the crank angle of the #2 cylinder. Similarly θ6 (protrusion 1
The crank pulse after detecting the cam pulse in 7c) is #
It shows the crank angle of 4 cylinders, and also the above θ6
When a cam pulse of θ4 (protrusion 17a) is detected after the cam pulse of , it can be determined that the crank pulse detected thereafter indicates the crank angle of the #1 cylinder.

Further, after the cam pulse is detected by the cam angle sensor 18, the crank pulse detected by the crank angle sensor 16 is determined to be the reference crank angle (θ1) of the corresponding cylinder.
).

The crank angle sensor 16 and the cam angle sensor 18 constitute crank angle detecting means, and by changing the cam pulse pattern, the cam angle sensor 18
The crank angle detection means may be configured only by the above.

On the other hand, a cooling water temperature sensor 20 faces a cooling water passage (not shown) forming a riser formed in the intake manifold 3 of the engine, and an exhaust pipe communicating with the exhaust port 2b of the cylinder head 2. O at 21
2 Sensor 22 is facing. Note that 23 is a catalytic converter, and 24 is a vehicle speed sensor.

(Circuit configuration of control device) On the other hand, reference numeral 31 is a control device consisting of a microcomputer, etc., and the control device 31 includes a CPU (central processing unit) 32, a ROM, etc.
33, RAM34, backup RAM (nonvolatile RAM)
M) 35 and I/O interface 36 are connected to each other via a bus line 37, and the constant voltage circuit 3
A predetermined stabilizing voltage is supplied from 8.

The constant voltage circuit 38 is connected to a battery 41 via a control relay 39, and a key switch 40 is turned on.
When the relay contact of the control relay 39 is closed, it supplies power for control to each part, and is directly connected to the battery 41, and provides backup power to the backup RAM 35 even when the key switch 40 is turned off. supply.

[0038] Also, the I/O interface 36
Each sensor 8, 9a, 16, 18, 20 is connected to the input port of
, 24, and the idle switch 9b are connected, and the positive terminal of the battery 41 is connected, and the terminal voltage is monitored. M
The PI 10 is connected to a warning means 43 such as an indicator lamp disposed on an instrument panel (not shown) or the like.

The ROM 33 stores control programs, fixed data, and the like. The fixed data includes a misfire determination level map MPΔNLEVE for each cylinder, which will be described later.
There is an initial set value ΔNSET of L, etc.

The RAM 34 also stores data obtained by processing the output signals of the sensors, data processed by the CPU 32, and the like. Furthermore, the backup RAM 35 is constantly powered regardless of the key switch 40, and the stored contents are not lost even if the key switch 40 is turned OFF and engine operation is stopped. Also, trouble data such as misfire data for each cylinder is stored.

This trouble data can be read by connecting a failure diagnosis serial monitor 45 to the failure diagnosis connector 44 connected to the output port of the I/O interface 36.

Furthermore, in the CPU 32, the ROM 3
According to the control program stored in 3, the above RAM
34. Based on various data stored in the backup RAM 35, calculate the fuel injection pulse width Ti for the MPI 10 for each cylinder.

In addition to general fuel injection control, the control device 31 individually determines misfire in each cylinder #i (i=1 to 4).

The basic concept of this misfire discrimination method for each cylinder is shown in FIG.
, will be explained according to FIG.

FIG. 8 shows engine rotational fluctuations.
For example, in the case of a 4-cylinder engine, the combustion stroke cylinder #i switches every 180°C according to the ignition order (for example, #1 → #3 → #2 → #4), so the combustion strokes do not overlap before and after that. , combustion stroke cylinder #
There is a so-called section in which no work is done by combustion, which is not affected by the combustion of each cylinder, between after the combustion of cylinder i ends and before the combustion of cylinder #i+1 in the next combustion stroke.

For example, as shown in FIG.
If the engine speed, which is the instantaneous momentum in the section where no work is done due to combustion of #4, is N#1 to N#4, respectively, if a misfire occurs in a certain cylinder #i,
Engine output after combustion drops sharply. In this embodiment, the combustion state of each cylinder #i and the engine speed N#i are
Noting that there is a very strong correlation between
NLEVEL to determine the misfire state of the cylinder #i.

That is, in this misfire determination method, first,
The average value of the engine speed N#i of the current combustion stroke cylinder #i and the engine speed N#i-2 of the cylinder #i-2 two combustion strokes ago (N#i-2+N#i)/2, and 1 Combustion state determination value ΔN#i-1 (ΔN#i-1←N#i-1
-(N#i+N#i-2)/2), and this discriminant value Δ
N#i-1 is compared with a misfire determination level ΔNLEVEL set according to the operating state of the cylinder #i-1,
The above combustion state determination value ΔN#i-1 is the misfire determination level ΔN
If it is lower than LEVEL, it is determined that there is a misfire.

Incidentally, in an actual engine, friction differs slightly from cylinder to cylinder due to finishing accuracy errors such as cylinder bore diameter and piston diameter, and manufacturing errors. When determining a misfire, the determination value ΔN#i-1 during normal combustion varies from engine to engine due to the influence of friction in each cylinder.

In this embodiment, in order to eliminate the variation factor for each cylinder due to friction, the misfire determination level ΔN
LEVEL is corrected by the discrimination value ΔN#i-1 at the time of fuel cut and is set as a learned variation value. A comparison value for determining a misfire is set based on the average value of engine speeds of adjacent cylinders in the firing order, and a misfire determination level ΔNLEVE is also set.
Misfire judgment level ΔNLE that eliminates the friction factor by learning and correcting L using the discrimination value ΔN#i−1 at the time of fuel cut
Since VEL is set for each cylinder, misfire conditions can be accurately detected not only when driving at a constant speed but also during acceleration.

[0050] In the following, each combustion stroke previous cylinder #i-
A method of calculating the combustion state determination value ΔN#i−1 of 1 will be specifically shown.

[0051]ΔN#1=N#1-(N#4+N#3)/
2ΔN#3=N#3-(N#1+N#2)/2ΔN#2
=N#2-(N#3+N#4)/2ΔN#4=N#4-
(N#2+N#1)/2 Also, below, a correlation equation between the engine rotational speed N in a section where no work is being done by combustion and the combustion state of the cylinder, that is, the indicated mean effective pressure Pi is shown.

[0052] First, the state in which the engine is rotating is expressed by the following formula: I・(2π/60)・(dN/dt)
=Ti-Tf... (1)

I: Moment of inertia

N: Engine rotation speed

Ti: Indicated torque

Tf: Friction torque, simplifying equation (1),

[0053]

[0054] Further, when expressed in terms of pressure,

[
0055

Pi: Indicated average effective pressure Pf: Friction loss effective pressure.

According to experiments, the engine speed N#i after combustion in each cylinder #i and the temporal change ΔT in the section in which this engine speed N#i is detected (for example, the section in FIGS. 6 and 8 ( dN of the above formula (3) based on θ2 - θ3 ) equivalent)
/dt, a very strong correlation was obtained.

Therefore, by determining the engine speed after combustion in each cylinder, it is possible to estimate the indicated mean effective pressure Pi, that is, the combustion state, and by comparing it with the engine speed of adjacent combustion stroke cylinders. , it is possible to set a determination value for determining whether or not there is a misfire in the relevant combustion stroke cylinder.

(Operation) Next, a specific cylinder-by-cylinder misfire determination procedure executed by the control device 31 will be described with reference to the flowcharts of FIGS. 1 to 3. Note that this flowchart is executed for each cylinder in synchronization with the rotation speed.

First, step (hereinafter abbreviated as "S") S1
01, the crank angle sensor 16 and the cam angle sensor 18
Combustion stroke cylinder #i (i=1, 3, 2, 4) based on the crank pulse and cam pulse output from
In S102, a crank pulse for detecting the BTDC θ2, θ3 outputted from the crank angle sensor 16 is determined by the interruption of the cam pulse, and in S103, the elapsed time between the crank pulses for detecting the BTDC θ2, θ3; The angle between θ2 and θ3 (θ2 −
Calculate the period f2,3 from θ3) (f2,3 ←d
t2,3/d(θ2 −θ3)). After that, S10
4, the engine rotation speed N# in the period from the above period f2,3 to the section where no work is being done by combustion in the current combustion stroke cylinder #i.
Calculate i (N#i←60/f2,3).

[0061] Then, in S105, the fuel cut flag FLA set in the fuel cut determination flowchart to be described later is
It is determined whether the GFC is in the set state (FLAGFC=1), and if FLAGFC=1 (fuel cut), the process proceeds to S106, and if FLAGFC=0 (fuel injection), the process proceeds to S107.

[0062] In S106, S115 or S133
Determine whether the fuel cut/cut release flag FLAG1 set in is set (FLAG1 = 1), and
If G1=0 (previous fuel injection), proceed to S108 and FL
If AG1=1 (previous fuel cut), the process advances to S109.

When the process advances to S108, since this is the first routine after fuel cut, the delay counter COUNT is set to the set value COUNTSET (in this embodiment, COUNT
After setting TSET = 2) (COUNT←CO
UNTSET), proceed to S115. The engine rotational speeds N#i-1 and N#i-2 of the cylinder #i-1 before the first combustion stroke and the cylinder #i-2 before the second combustion stroke immediately after the fuel cut are data at the time of fuel injection, and will be described later. If read in S111, friction for each cylinder cannot be detected. Therefore, after fuel cut in S111, at least 2
A delay count is set in S108 in order to read out engine rotational speeds N#i-1 and N#i-2 for combustion stroke cylinder #i after the sequential combustion stroke.

On the other hand, in S106, the previous fuel cut (FLA
G1=1) and proceeds to S109, it is determined whether the delay counter COUNT is 0, and the COUNT
If ≠0, after subtracting in S110 (COUNT←CO
UNT-1), proceed to S115. Further, if it is determined in S109 that COUNT=0, the process proceeds to S111 and learning of the misfire determination level ΔNLEVEL is executed.

First, in S111, the engine rotational speed N#i- in the section where no work is being done by combustion in the cylinder #i-1 preceding one combustion stroke, which was set in the previous and two previous routines and stored at a predetermined address in the RAM 34, is determined. The engine rotational speed N#i-2 in the section where no work is being done by combustion in the cylinder #i-2 before the first and second combustion strokes are read out. As shown in FIG. 8, when the ignition order is #1 → #3 → #2 → #4, if the current combustion stroke cylinder #i is #3, the cylinder #i-1 one combustion stroke before is #1, 2 combustion stroke front cylinder #i-2
becomes #4.

Next, in S112, the current combustion stroke cylinder #i
The average value (N From the difference between #i+N#i-2)/2 and the engine speed N#i-1 in the section where the cylinder #i-1 before one combustion stroke is not doing work due to combustion, the cylinder #i after one combustion stroke is determined.
-1, the combustion state discrimination value ΔN#i-1 is determined.

ΔN#i-1←N#i-1-{(N#i+N#i-2)
/2} Then, in S113, one combustion stroke previous cylinder #i-1
The engine rotation speed N#i-1 of the cylinder #i-1 is stored in the backup RAM 35 as a parameter.
The misfire determination level ΔNLEVEL is set from the misfire determination level map MPΔNLEVEL.

As shown in FIG. 9, the misfire determination level map MPΔNLEVEL is a two-dimensional map using the engine speed as a parameter, and is provided for each cylinder, and the misfire determination level ΔNLEVEL is stored in each region. There is. Note that the initial value of the misfire judgment level ΔNLEVEL is ΔNSET, which was determined in advance through experiments.
is set.

Then, in S114, the combustion state determination value Δ
Based on N#i-1 and the above-mentioned misfire judgment level ΔNLEVEL, the misfire judgment level is calculated from the weighted average of the weight r shown in the following formula, and the misfire judgment level is calculated at the corresponding address of the misfire judgment level map MPΔNLEVEL of the cylinder #i-1 one combustion stroke before. The stored data is updated and the process advances to S115.

ΔNLEVEL ←{(2r −1)×Δ
NLEVEL + (ΔN#i-1+k)}/2r k:
The amount of correction as a margin determined in advance through experiments etc.
Since the friction exerted by the engine differs from cylinder to cylinder,
It is necessary to learn the misfire determination level for each operating region for each cylinder.

[0071] The above S108, S110, or S1
14 to S115, the fuel cut/cut release flag FLAG1 is set (FLAG1←1) and the routine is exited.

The misfire determination level ΔNLEVEL updated in S114 is, for example, ΔNLEVEL←{(2r −1)×Δ
NLEVEL + (ΔN#i-1×I)}/2r I:
The correction ratio as a margin set in advance through experiments etc.
For example, I = 1.1), and if weighted average is not used, ΔNLE
VEL ←ΔN#i-1+k Alternatively, ΔNLEVEL ←ΔN#i-1×I may be used.

On the other hand, in S105 above, the fuel cut flag F
When it is determined that LAGFC is in the reset state (FLAGFC=0, fuel injection) and the process proceeds to S107, the fuel cut/cut release flag FLAG1 is set (FLAG1=1).
), and FLAG1 = 1 (previous fuel cut).
If so, the process advances to S116, and if FLAG1=0 (previous fuel injection), the process advances to S117.

When the process advances to S116, since this is the first routine after restarting fuel injection, the delay counter COUNT is set to the set value COUNTSET (in this embodiment, COUNT
After setting TSET = 2) (COUNT←CO
UNTSET), proceed to S133. Engine rotational speeds N#i-1, N#i-2 of the cylinder #i-1 before the first combustion stroke and the cylinder #i-2 before the second combustion stroke immediately after restarting fuel injection
Since this is data at the time of fuel cut, if a misfire is determined based on the value read in S120, which will be described later, an erroneous determination will occur. Therefore, in S120, the engine rotation speed N# at least in the second combustion stroke cylinder #i after restarting fuel injection.
In order to read out i-1 and N#i-2, a delay count is set in S116.

On the other hand, in S107, the previous fuel injection (FL
When it is determined that AG1=0) and the process proceeds to S117, it is determined whether the delay counter COUNT is 0, and COUNT is
If T≠0, after subtracting in S118 (COUNT←C
OUNT-1), proceed to S133. In addition, the above S117
If it is determined that COUNT=0 in step S119, a misfire determination for cylinder #i-1 one combustion stroke before is executed.

First, in S119, the combustion stroke cylinder #i
Count up the number of calculation cycles C#i1 (C
#i1 ←C#i+1).

Next, in S120, the engine rotational speed N#i-1 in the section where no work is being done by combustion in the cylinder #i-1 preceding one combustion stroke, which was set in the previous and two previous routines and stored at a predetermined address in the RAM 34, is determined. and the engine rotation speed N#i-2 in the section where no work is being done by combustion in the cylinder #i-2 before the 2nd combustion stroke, and in S121, the section where no work is being done by combustion in the cylinder #i in the current combustion stroke is read out. The engine speed N#i in the section where no work is done by combustion in the 2nd combustion stroke previous cylinder #i-2
-2 and the average value (N#i+N#i-2)/2 and the above 1
The combustion state discrimination value ΔN#i-1 of the cylinder #i-1 before the combustion stroke is determined from the difference from the engine rotation speed N#i-1 in the section where the cylinder #i-1 before the combustion stroke is not doing work due to combustion. demand.

[0078]ΔN#i-1←N#i-1-(N#i+N
#i-2)/2 Thereafter, in S122, a misfire determination level ΔNLEVEL is set from the misfire determination level map MPΔNLEVEL using the engine rotation speed N#i-1 determined in the previous routine as a parameter.

As shown in FIG. 10, the combustion state discrimination value ΔN
#i-1 shows a relatively large amount of variation during acceleration, and this amount of variation varies depending on the operating range of the engine, but since the friction exerted by the engine is excluded from the misfire judgment factor, it is a delicate area for misfire judgment. Even if there is, it can be determined accurately.

[0080] Then, in S123, the misfire determination level ΔNLEVEL and the cylinder #i-1 before the first combustion stroke are determined.
If it is determined that the combustion state determination value ΔN#i-1 is lower than the misfire determination level ΔNLEVE (ΔN#i-1<ΔNLEVEL) (see FIG. 10), It is determined that there is a misfire and the process proceeds to S124, and if ΔN#i-1≧ΔNLEVEL, it is determined that the combustion is normal and the process proceeds to S125.

When it is determined that a misfire has occurred and the process proceeds to S124, the number of misfires by cylinder C (#i-1) for the cylinder #i-1 one combustion stroke before is determined.
2 is counted up and (C(#i-1)2 ←C(#
i-1)2+1), proceed to S125.

[0082] Then, in S125, the current combustion stroke cylinder #i
The number of calculation cycles C#i1 and the preset number of sampling cycles C#i1SET (for example, 100cyc
le), and if the number of calculation cycles C#i1 does not reach the number of sampling cycles C#i1SET (
C#i1<C#i1SET), jump to S133,
Also, when the number of calculation cycles C#i1 reaches the number of sampling cycles C#i1SET (C#i1 ≧C#i
1SET), proceed to S126, and enter the above calculation cycle number C#.
Clear i1 (C#i1 ← 0).

Next, in S127, the previous sample stored at a predetermined address in the RAM 34 is


The average number of misfires by cylinder C (#i
−

1)
2(-1) is read out, and in S128, this average number of misfires for each cylinder C(#i-1)2(-1) and 1 counted in the current sampling cycle number C#i1SET are read out.
Cylinder #i-1 before combustion stroke


Number of misfires by cylinder C(#i-1)2
Based on the current average number of misfires for each cylinder C(#i-1)
2 is obtained from the weighted average of the weights r shown in the following equation.

[0084]

C(#i-1)2 ←{(2r -1
)×C(#i-1)2(-1) +C(#i-1)2
}/2r


By calculating the average number of misfires C(#i-1)2 for each cylinder by weighted average,
It is possible to correct misfire detection errors and temporary misfire misjudgments due to sudden combustion fluctuations.

Thereafter, in S129, the average number of misfires for each cylinder C(#i-1)2 is cleared (C(#i-1)2).
←0), and the previous sample stored at a predetermined address in the RAM 34 in S130.

The average number of misfires for each cylinder calculated in the ring cycle C(#i-1)2(-1) was calculated this time.

Update the average number of misfires by cylinder C(#i-1)2 (C(#i
-1)2(-1)←C(#i-1)2).


Then, in S131, the current average number of misfires for each cylinder C (#i-1
)2 and the preset misfire


Abnormality judgment standard number of times C (#i-1)
2SET, C(#i-1)2 >C(#i-
1) 2SET, Sunawa


The average number of misfires by cylinder C (#i-1)
2 exceeds the misfire abnormality determination reference number C (#i-1) 2SET, it is determined that the cylinder #i-1 one combustion stroke before has a misfire abnormality, and the process proceeds to S132, where the backup RA
The misfire abnormality data of cylinder #i-1 one combustion stroke before is stored in a predetermined address of M35, and the warning means 43 such as an indicator lamp is lit to notify the driver of the misfire.

A fire abnormality warning is given and the process proceeds to S133. On the other hand, if it is determined that C(#i-1)2 ≦C(#i-1)2SET, it is determined that a misfire abnormality has not yet occurred in cylinder #i-1 one combustion stroke before, and the process directly proceeds to S133. .

The misfire abnormality data for cylinder #i-1 one combustion stroke prior, stored in the backup RAM (storage means) 35, can be read by connecting the serial monitor 45 to the control device 31 at a dealer's service station or the like. By checking the serial monitor 4, you can determine which cylinder is misfiring.
5, the misfire abnormality data stored in the backup RAM 35 can be cleared.

[0087] Then, the above S116, S118, S12
5. When proceeding from S131 or S132 to S133, the fuel cut/cut release flag FLAG1 is reset (FLAG1←0) and the routine exits.

Next, an example of the fuel cut determination procedure will be explained according to the flowchart of FIG. 4. First, in S201, the vehicle speed S and the fuel cut condition determination reference vehicle speed S0 (for example, S
0 = 15km/h), and if S0 ≦S, S
The process advances to S202, and if S0>S, the process advances to S205.

When the process advances to S202, the engine rotation speed N and the fuel cut condition determination reference engine rotation speed N0 (for example,
(N0 = 1500 rpm), and if N0≦N, the process advances to S203, and if N0>N, the process advances to S205.

When the process advances to S203, it is determined whether the idle switch 9b is ON, and if the idle switch 9b is ON.
If the throttle valve 5a is fully closed, the process proceeds to S204, and if it is OFF (the throttle is open), the process proceeds to S205.

Fuel cut condition satisfied (S0 ≦S, N0
≦N and the idle switch is ON), and S2
When the program proceeds to 04, the fuel cut flag FLAGFC is set (FLAGFC←1) and the routine is exited.

[0092] Also, the fuel cut condition is not satisfied (S0 > S
, N0>N, or the idle switch is OFF), and the process proceeds to S205, the fuel cut flag FLAG
Reset the FC (FLAGFC←0) and exit the routine.

(Second Embodiment) FIGS. 11 and 12 are flowcharts showing a cylinder-by-cylinder misfire determination procedure according to a second embodiment of the present invention.

In this embodiment, the number of misfires is sequentially stored for each cylinder, and when the number of misfires reaches the maximum count, the maximum number of misfires is fixed and stored.

First, in S101, the combustion stroke cylinder #i is determined, and then in S301, the misfire maximum count flag FLAG#i-1 for the cylinder #i-1 one combustion stroke before is set (FLAG#i-1= 1) or reset state (FLA
G#i-1=0). Set state (FLAG
#i-1=1), the process advances to S302, and the warning means 43
An indicator lamp (see FIG. 5) is turned on to warn the driver of the misfire abnormality, and the routine is exited.

On the other hand, the misfire maximum count flag FL
AG#i-1 is in reset state (FLAG#i-1=0)
If it is determined that
The above-mentioned first embodiment (Fig. 1, Fig. 2)
(S106 to S115 are shown in Figure 2).
(see the same figure).

[0097] Then, in the above S123, the cylinder #i-1 one combustion stroke before misfires (ΔN#i-1<ΔNLEVEL).
If it is determined that the misfire has occurred, the process proceeds to S303, where an indicator lamp or the like is turned on for a short period of time to warn the driver that a misfire has occurred, and then the process proceeds to S124.

By recognizing the frequency with which indicator lamps and the like are lit, the driver can grasp the engine misfire situation, that is, under what engine operating conditions misfires are likely to occur.

[0099] Also, in S123, normal combustion (ΔN#i-1
≧ΔNLEVEL), the process advances to S307,
Reset the misfire maximum count flag FLAG#i-1 for cylinder #i-1 before one combustion stroke (FLAG#i-1←0
), the process advances to S133.

[0100] Then, in S124, one combustion stroke previous cylinder #
After counting up the number of misfires C(#i-1)2 of i-1 (C(#i-1)2 ←C(#i-1)2 +1), this counted up value C(#i-1)2 1) Store 2 at a predetermined address in the backup RAM 35.

[0101] At a dealer's service station, etc., a serial monitor 45 is connected to the control device 31 via a fault diagnosis connector 44 and the backup RAM 3 is
The number of misfires for each cylinder stored in 5 C(#i-1)2
Read the data and refer to the manual to determine the misfire situation.

[0102] After that, in S304, one combustion stroke previous cylinder #
Compare the number of misfires C(#i-1)2 of i-1 with a preset maximum count number FFH (however, it can be set arbitrarily depending on the capacity of the microcomputer), and calculate C(#i-1)
If 2=FFH, the process advances to S305, and C(#i-1
)2 <FFH, the process advances to S307.

[0103] The number of misfires C(#i-1)2 has reached the maximum count number FFH (C(#i-1)2=F
FH) and proceeds to step S305, the number of misfires C(#i-1)2 stored at a predetermined address in the backup RAM 35 is held at the maximum count number FFH, and in S306, the number of misfires stored in the predetermined address of the backup RAM 35 is held at the maximum count number FFH. #i-1
The misfire maximum count flag FLAG#i-1 is set (FLAG#i-1←1), and the routine is exited.

(Third Embodiment) FIGS. 13 and 14 are flowcharts showing a cylinder-by-cylinder misfire determination procedure according to a third embodiment of the present invention.

In this embodiment, the combustion stroke cylinder #i
Engine rotation speed N in the section where no work is done by combustion of
A misfire determination level map for the cylinder #i based on the difference rotational speed ΔN#i between #i and the engine rotational speed N#i-1 in the section where the cylinder #i-1 one combustion stroke before is not doing work due to combustion. The misfire determination level ΔNLEVEL stored in the relevant operating region of MPΔNLEVEL is corrected and learned.

[0106] This will be explained below according to the flowchart. Note that routines that are the same as those in the first embodiment are designated by the same reference numerals, and a description thereof will be omitted.

First, steps S101 to S110 are carried out in the same manner as in the first embodiment (see FIGS. 1 and 13), and S401
Learning of the misfire determination level ΔNLEVEL is executed below.

[0108] In S401, set in the previous routine, R
1 combustion stroke previous cylinder # stored in a predetermined address of AM34
The engine rotation speed N#i-1 in the section where no work is being done by combustion of cylinder #i in the current combustion stroke is read out, and in S402, the engine rotation speed N#i in the section where no work is being done by combustion in the current combustion stroke cylinder #i is read out. The differential rotational speed ΔN#i of the cylinder #i in the current combustion stroke is determined from the difference from the engine rotational speed N#i-1 (ΔN#i←N#i-N#i-1).

[0109] Also, in S403, the engine rotation speed N#
Misfire determination level map MPΔ of the relevant cylinder #i stored in the backup RAM 35 with i as a parameter
Misfire judgment level ΔNL from NLEVEL (see Figure 9)
Set EVEL.

[0110] Then, in S404, the differential rotational speed ΔN#
Based on i and the above misfire judgment level ΔNLEVEL, the misfire judgment level ΔN is determined from the weighted average of the weight r shown in the following formula.
LEVEL is determined, the data stored in the corresponding address of the misfire determination level map MPΔNLEVEL for the current combustion stroke cylinder #i is updated, and the process proceeds to S115.

ΔNLEVEL ←{(2r −1)
×ΔNLEVEL +(ΔN#i+k)}/2r On the other hand, the same procedure as in the first embodiment is executed from S116 to S119 (see FIGS. 1 and 14), and the misfire determination for the current combustion stroke cylinder #i is performed at S405 and below. Execute.

[0112] First, in S405, the cylinder #i-
1 is read out, and in S406, the combustion state determination value ΔN#i of the current combustion stroke cylinder #i is determined based on the engine rotation speed N#i of the current combustion stroke cylinder #i and the engine rotation speed N Determine from the difference from #i-1 (ΔN#i←N
#i-N#i-1).

[0113] Also, in S407, the engine rotation speed N#
Misfire determination level map MPΔNL with i as a parameter
Set the misfire judgment level ΔNLEVEL from EVEL.

Then, in S408, the misfire determination level ΔNLEVEL and the combustion state determination value ΔN#i are compared, and if ΔN#i<ΔNLEVEL (see FIG. 10), it is determined that a misfire has occurred and the process proceeds to S409. ΔN#i≧ΔN
In the case of LEVEL, it is determined that combustion is normal and the process proceeds to S125.

When it is determined that a misfire has occurred and the process proceeds to S409, the number of cylinder-specific misfires C#i2 for the cylinder #i in the current combustion stroke is counted up (C#i2←C#i2+1), and the process proceeds to S125.

[0116] After that, S125, S126 to S410
, the cylinder-by-cylinder average of the current combustion stroke cylinder #i calculated in the previous sampling period stored in a predetermined address of the RAM 34 is displayed.

The average number of misfires C#i2(-1) is read out, and in S411, this average number of misfires for each cylinder C#i2(-1) and the current combustion stroke number counted in the current sampling cycle number C#i1SET are calculated.

Based on the number of misfires by cylinder C#i2 of cylinder #i, the current average number of misfires by cylinder C#i2 is determined from the weighted average of the weight r shown in the following equation.

[0117]

C#i2 ←{(2r -1)×C#i2(-1) +
C#i2 }/2r Then, in S412, the number of misfires by cylinder C#i2 is cleared (C#i2 ← 0), and in S413, the previous sampling period stored in the predetermined address of the RAM 34 is cleared.

The average number of misfires by cylinder C#i calculated in
2(-1) is the average number of misfires for each cylinder calculated this time.


Update with the number C#i2 (C#i2(-1) ←C#i2).

[0118]

Then, in S414, the current average number of misfires by cylinder C#i
2 and preset misfire abnormality


Compare the judgment standard number of times C#i2SET and find that C#i2 >C#i2SET, that is, the average loss for each cylinder.


The number of fires C#i2 is the reference number of misfire abnormality judgments C#i2
If it exceeds SET, the current combustion stroke


It is determined that cylinder #i has a misfire abnormality, and the process proceeds to S132, while C#i2≦C
If #i2SET is determined, it is determined that the misfire abnormality has not yet occurred in the current combustion stroke cylinder #i, and the S1
33 (see Figure 1).

[0119] The present invention is not limited to the above-mentioned embodiments; for example, the combustion state discrimination value may be determined by using angular acceleration, period, angular velocity, etc., as the work due to combustion in each cylinder as the momentum of the specified section. You can do it like this.

[0120]

[Effects of the Invention] As explained above, according to the present invention,
The misfire condition is determined without including the combustion condition factors of other cylinders.
Excellent effects such as accurate detection can be achieved without being affected by combustion variations between cylinders, manufacturing variations from engine to engine, and engine friction.

[Brief explanation of the drawing]

FIG. 1 to FIG. 3 are flowcharts showing a misfire determination procedure for each cylinder.

[Figure 2] Same as above

[Figure 3] Same as above

FIG. 4 is a flowchart showing a fuel cut determination procedure.

FIG. 5 is a schematic diagram of the engine control system.

FIG. 6 is a front view of the crank rotor and crank angle sensor.

FIG. 7 is a front view of a cam rotor and a cam angle sensor.

FIG. 8 is a time chart showing cylinder pressure fluctuations, crank pulses, cam pulses, and engine rotation fluctuations.

FIG. 9 is a conceptual diagram of a misfire determination level map.

FIG. 10 is a conceptual diagram showing a basic misfire determination method for each cylinder.

FIG. 11 is a flowchart showing a cylinder-by-cylinder misfire determination procedure according to a second embodiment of the present invention.

[Figure 12] Same as above

FIG. 13 is a flowchart showing a cylinder-by-cylinder misfire determination procedure according to a third embodiment of the present invention.

[Figure 14] Same as above

[Explanation of symbols]

Claims (2)

[Claims] Claim 1: 1 from the difference between the average value of the momentum in a predetermined section of the cylinder before two combustion strokes and the momentum in a predetermined section of the cylinder in the current combustion stroke at the time of fuel cut, and the momentum in the predetermined section of the cylinder before one combustion stroke. The combustion state discrimination value of the cylinder before the combustion stroke is calculated, and the misfire judgment level stored at a predetermined address of the misfire judgment level map of the cylinder before one combustion stroke is read out using the momentum of the cylinder before one combustion stroke as a parameter. It is updated with the value corrected by the state discrimination value, and then the average value of the momentum of the predetermined section of the cylinder two combustion strokes prior to fuel injection and the momentum of the predetermined section of the current combustion stroke cylinder, and the predetermined section of the cylinder one combustion stroke prior. The combustion state discrimination value of the cylinder 1 combustion stroke before is calculated from the difference between the momentum of the cylinder 1 combustion stroke before, and the combustion state discrimination value of the cylinder 1 combustion stroke before and the above-mentioned momentum of the cylinder 1 combustion stroke before are read as parameters. A misfire determination method for each cylinder in an engine, characterized in that a misfire state of a cylinder one combustion stroke before is determined by comparing the misfire determination level with a misfire determination level stored at a predetermined address of a misfire determination level map. 2. A combustion state determination value for the current combustion stroke cylinder is determined from the difference between the momentum in a predetermined section of the current combustion stroke cylinder at the time of fuel cut and the momentum in a predetermined section of the cylinder one combustion stroke before; The misfire determination level stored in a predetermined address of the misfire determination level map for the cylinder in the current combustion stroke read out using the momentum as a parameter is updated with a value corrected by the combustion state determination value, and then the current combustion stroke during fuel injection is updated. The combustion state discrimination value of the current combustion stroke cylinder is calculated from the difference between the momentum of the cylinder in a predetermined section and the momentum of the cylinder one combustion stroke previous. A misfire state of a cylinder in a current combustion stroke is determined by comparing a misfire determination level map stored at a predetermined address of a misfire determination level map of a cylinder in a current combustion stroke read out using momentum as a parameter. Misfire detection method.