JP2821251B2 - Engine idle control device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an engine idling control device for reducing fluctuations in engine rotation during idling.

[Related Art] In recent years, the idle speed of an engine has tended to be kept low due to demands for improved fuel efficiency and quietness. However, when the idle speed is reduced, the rotation tends to become unstable, and comfort and start-up are reduced. Performance is significantly impaired.

Generally, it is considered that the rotation fluctuation is caused by a variation in the combustion state of each cylinder, and the variation in the combustion state is caused by a complicated intake pipe shape, a non-uniform intake distribution ratio due to intake interference of the cylinder pipe, and the like. A slight difference in combustion temperature between the cylinders caused by the cooling route, a variation in the combustion chamber volume of each cylinder, manufacturing variations in piston shape, etc., and a difference in the air-fuel ratio of each cylinder resulting from a difference in fuel injection amount due to a manufacturing error of the injector, etc. It is thought to occur due to slight variations, etc.

Therefore, by controlling the combustion state of each cylinder substantially uniformly, the idle rotation speed can be made smooth.

Means for uniformly controlling the combustion state of each cylinder include fuel injection control and ignition timing control.However, in fuel injection control, control must be performed within the stoichiometric air-fuel ratio range, and the control range is greatly limited. Is done.

On the other hand, in the ignition timing control, there is a difficulty in controllability because the combustion state of the engine fluctuates remarkably by a slight advance or retard correction.

Therefore, for example, as disclosed in Japanese Patent Application Laid-Open No. 2-64252, there is an apparatus that controls both fuel injection control and ignition timing control simultaneously.

That is, the engine rotation fluctuation is detected at every predetermined crank angle, and based on the rotation fluctuation, the fuel injection amount and the ignition timing of the cylinder having a large fluctuation difference are corrected, and the fuel injection amount and the ignition timing of the other cylinders are corrected. The control is performed such that the sum of the fuel injection amount and the correction amount of the ignition timing of all cylinders is always zero.

[Problems to be Solved by the Invention] By the way, when simultaneously controlling the fuel injection amount and the ignition timing in one engine, after setting either one, an effective air-fuel ratio is required unless a region is set based on this set value. If the controllability and the ignition timing controllability cannot be obtained, for example, when the increase correction and the advance correction are simultaneously performed on a cylinder having a negative rotation fluctuation difference (cylinder in a bad combustion state), the cylinder and other The difference between the air-fuel ratio and the torque fluctuation between the cylinder and the cylinder becomes large, and the feeling is deteriorated.

[Object of the Invention] The present invention has been made in view of the above circumstances, and provides an engine idle control device that can uniformly control the combustion state of each cylinder without deteriorating the feeling during idling. It is intended to be.

[Means for Solving the Problems] To achieve the above object, an engine idle control device according to the present invention comprises, as shown in FIG. 1, an operating state detecting means M1 for detecting an operating state of an engine, Crank angle detection means M2 for outputting a signal capable of determining a certain cylinder and a signal corresponding to a predetermined crank angle, idle determination means M3 for determining an idle state of the engine from an engine operation state, and when the engine is in an idle state, A combustion state detecting means M4 for detecting an engine rotation fluctuation difference for each predetermined crank angle, and comparing the engine rotation fluctuation difference with an allowable rotation fluctuation limit to detect a combustion state of a cylinder in a combustion stroke.
When the rotation fluctuation difference of the cylinder is out of the permissible rotation limit when the engine is idling, the injection correction amount of each cylinder is calculated by summing the absolute value of the rotation fluctuation difference of each cylinder and the rotation of the cylinder. If the injection correction amount of the cylinder exceeds the allowable upper limit, the injection correction amount of the cylinder is reduced by a set amount and the injection correction amount of the other cylinder is reduced. If the injection correction amount of the cylinder is lower than the allowable lower limit, the injection correction amount of the cylinder is increased by a set amount, and the injection correction amount of the other cylinder is distributed and reduced by the increase amount. An injection correction amount setting means M5 for calculating the basic injection amount based on the operating state of the engine, and an injection amount calculation means M6 for correcting the basic injection amount with the injection correction amount for each cylinder when the engine shifts to the idle state; Based on the output of the injection amount calculation means The fuel supply means M7 for supplying fuel to the corresponding cylinder, and when the injection correction amount of the cylinder exceeds an allowable upper limit when the engine is in an idle state, the ignition timing correction amount of the cylinder is advanced by a predetermined angle. If the injection correction amount of the cylinder is lower than the permissible lower limit value, the ignition timing correction amount setting means M8 sets the ignition timing correction amount of the cylinder delayed by a predetermined angle, and the basic ignition based on the operating state of the engine. The ignition timing calculating means M9 calculates the timing, and shifts to the idle state to correct the basic ignition timing for each cylinder with the ignition timing correction amount.
And ignition means M10 for igniting the air-fuel mixture of the corresponding cylinder based on the output of the ignition timing calculation means.

[Operation] In the present invention, during the idle state, the combustion state of the cylinder in the combustion stroke is detected by comparing the engine rotation fluctuation difference for each predetermined crank angle with the rotation fluctuation allowable limit. And
If the rotation fluctuation difference of the cylinder is out of the rotation fluctuation allowable limit, the injection correction amount of each cylinder is determined according to the ratio between the sum of the absolute values of the rotation fluctuation difference of each cylinder and the rotation fluctuation difference of the cylinder. Correct each. Further, if the injection correction amount of the cylinder exceeds the allowable upper limit, the ignition timing correction amount of the cylinder is advanced by a predetermined angle, and the injection correction amount of the cylinder is reduced by a set amount, and the injection correction amount of the other cylinder is reduced. The injection correction amount is distributed and increased by the reduced amount. Further, when the injection correction amount of the cylinder is lower than the permissible lower limit value, the ignition timing correction amount of the cylinder is retarded by a predetermined angle, and the injection correction amount of the cylinder is increased by a set amount, and the injection correction amount of the other cylinder is increased. Distribute and reduce the amount by the increment.

In the idle state, the basic injection amount and the basic ignition timing are corrected by the injection correction amount and the ignition timing correction for each cylinder.

Therefore, the control is widened and the controllability is greatly improved without causing overlean or overrich in each cylinder, and the injection correction amount is corrected in accordance with the rotation fluctuation difference of each cylinder. The rotation fluctuation difference for each cylinder is eliminated, and a good feeling can be obtained.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

2 to 11 show a first embodiment of the present invention, FIG. 2 is an overall schematic diagram of an engine control system, FIG. 3 is a front view of a crank rotor and a crank angle sensor, and FIG. Front view of the cam angle sensor, FIG. 5 is a conceptual diagram of a basic ignition timing map, FIG. 6 is a time chart of cylinder pressure fluctuation, clan pulse, cam pulse, and engine speed, and FIG. 7 is an average difference for each cylinder. FIG. 8 is a flowchart showing a cylinder-by-cylinder fuel injection control procedure, FIG. 9 is a flowchart showing a cylinder-by-cylinder ignition timing control procedure, FIG. 10 is a cylinder-by-cylinder average differential rotation speed, invalid by cylinder FIG. 11 is a time chart showing an injection pulse width, and FIG. 11 is a time chart showing a cylinder-by-cylinder ignition timing correction learning value.

(Structure) Reference numeral 1 in FIG. 2 denotes an engine, which indicates a four-cylinder horizontally opposed engine in the figure. An intake manifold 3 communicates with an intake port 2a formed in a cylinder head 2 of the engine 1. A throttle chamber 5 is communicated upstream of the intake manifold 3 via an air chamber 4, and intake air is provided upstream of the throttle chamber 5. An air cleaner 7 is attached via a pipe 6.

An intake air amount sensor (hot wire air flow meter in the figure) 8 is interposed immediately downstream of the air cleaner 7 in the intake pipe 6, and a throttle valve 5 a provided in the throttle chamber 5 is provided with a throttle valve. An opening sensor 9a and an idle switch 9b for detecting the full closing of the throttle valve are connected to each other.

An injector (fuel supply means) is provided immediately upstream of each intake port 2a of each cylinder of the intake manifold 3.
Ten are arranged. Further, an ignition plug 11 for exposing the ignition portion at the tip to the combustion chamber is attached to each cylinder of the cylinder head 2. The ignition coil 11a is integrally attached to the ignition plug 11.

Further, the injector 10 is communicated with a fuel tank 13 via a fuel passage 12, and a fuel pump 14 is provided in the fuel passage 12.

Further, a crank rotor 15 is mounted on a crank shaft 1b of the engine 1 and a crank angle sensor including an electromagnetic pickup for detecting a protrusion (or a slit) corresponding to a predetermined crank angle is provided on the outer periphery of the crank rotor 15. A cam rotor 17 is connected to a camshaft 1c that rotates 1/2 of the crankshaft 1b, and a cam angle sensor 18 is provided on the outer periphery of the cam rotor 17.

As shown in FIG. 3, protrusions 15a, 15b, and 15c are formed on the outer periphery of the crank rotor 15. These projections 15a, 15
b and 15c are formed at positions before the compression top dead center (BTDC) θ1, θ2 and θ3 of each cylinder, and the period f1.2 (here, f = 1 / ω) is determined from the passage time between the protrusions 15a and 15b. ω: angular velocity), and the period f2.3 is calculated from the passing time between the protrusions 15b and 15c. Further, the projection 15b indicates a reference crank angle when setting the ignition timing ADV.

As shown in FIG. 6, the crank angle B of the projections 15b, 15c
TDC θ2, θ3 is the ignition timing (time) ADV during idling operation
Is set before and after. Generally, the ignition timing at the time of the idling operation is around 20 ° C. BTDC, and even if the ignition is performed at this crank angle, the combustion pressure does not increase rapidly until about 10 ° C. thereafter.

Further, as shown in FIG. 6, in the embodiment, the valve opening timing of the exhaust valve of each cylinder is set slightly retarded with respect to each BTDC θ2 of the ignition reference crank of the next combustion cylinder. Since the combustion pressure immediately after the opening of the exhaust valve is rapidly reduced, the combustion pressure has almost no effect at the crank angle BTDCθ3.

Therefore, the crank angle θ3 of the projection 15c is
If it is set on the more advanced side than ℃ A, the section between the crank angles BTDC θ2, θ3 of the projections 15b, 15c is hardly affected by the combustion between the cylinders, that is, the combustion stroke cylinder and the next combustion stroke This is a section where there is no work due to combustion with the cylinder.

As shown in FIG. 4, on the outer periphery of the cam rotor 17, projections (or slits) 17a, 17b, 17c for cylinder discrimination are provided.
Are formed. The projection 17a is formed at the position of θ4 after the top dead center (ATDC) of the # 3 and # 4 cylinders, and the projection 17b is composed of three projections, and the first projection is the top dead center of the # 1 cylinder. A point (ATDC) is formed at the position of θ5, and the projection 17c is composed of two projections.
It is formed at a position θ6 after the compression top dead center (ATDC) of the cylinder.

In the embodiment shown in the figure, θ1 = 97 ° C., θ2 = 65 ° C.,
θ3 = 10 ° C, θ4 = 20 ° A, θ5 = 5 ° A, θ6 = 20
° C, θ (2-3) = 55 ° C.
As shown in FIG. 6, for example, the cam angle sensor 18 detects θ
When the cam pulse of No. 5 (projection 17b) 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.

After the cam pulse of θ5, θ4 (projection 17a)
Is detected, it can be determined that the subsequent crank pulse detected by the crank angle sensor 16 indicates the crank angle of the # 2 cylinder. Similarly, θ6 (projection
17c) The crank pulse after detecting the cam pulse is #
It indicates the crank angle of the four cylinders.
When the cam pulse of θ4 (projection 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, it can be determined that the crank pulse detected by the crank angle sensor 16 indicates the reference crank angle (θ1) of the corresponding cylinder.

Note that the crank angle sensor 16 and the cam angle sensor 18 constitute crank angle detecting means, and the crank angle detecting means may be constituted only by the cam angle sensor 18 by changing a cam pulse pattern.

On the other hand, a knock sensor 19 for detecting knock from the vibration of the engine body 1 is fixed to the engine body 1,
A cooling water temperature sensor 20 faces a riser formed in the intake manifold 3 to a cooling water passage (not shown).

An O2 sensor 22 faces an exhaust pipe 21 communicating with the exhaust port 2b of the cylinder head 2. The sign
23 is a catalytic converter, 24 is a vehicle speed sensor, and 25 is an igniter.

(Circuit Configuration of Control Device) On the other hand, reference numeral 31 denotes a control device including a microcomputer or the like, and a CPU (central processing unit) 3
2, ROM33, RAM34, backup RAM (non-volatile RAM) 3
5, and the I / O interface 36 are connected to each other via the bus line 37, and a predetermined stabilized voltage is supplied from the constant voltage circuit 38.

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

In addition, to the input port of the I / O interface 36,
Each sensor 8, 9a, 16, 18, 19, 20, 22, 24, and the idle switch 9b are connected, the plus terminal of the battery 41 is connected, the terminal voltage is monitored, and the I I / O interface 36 output port with spark plug 1
1 is connected via an igniter 25, and the injector 10 is connected via a drive circuit.

The ROM 33 stores a control program, fixed data, and the like. The fixed data includes a basic ignition timing map MPθBASE described later, an initial set value, and the like.

The RAM 34 stores data obtained by processing output signals of the sensors, data processed by the CPU 32, and the like. Further, the backup RAM 35 is always supplied with power irrespective of the key switch 40. Even if the key switch 40 is turned off and the operation of the engine is stopped, the stored contents are not lost. And ignition timing correction learning value data.

Further, the CPU 32 calculates a fuel injection pulse width Ti for the injector 10 or an ignition timing ADV for each cylinder based on various data stored in the RAM 34 and the backup RAM 35 according to a control program stored in the ROM 33.

The idling control function of the control device 31 includes an idling determining means for judging an idling state of the engine from an operating state of the engine and, when the engine is in an idling state, detecting an engine rotation fluctuation difference for each predetermined crank angle. Combustion state detection means for comparing the engine rotation fluctuation difference with the rotation fluctuation allowable limit to detect the combustion state of the cylinder in the combustion stroke; and when the engine is idle, the rotation fluctuation difference of the cylinder deviates from the rotation allowable limit. In this case, the injection correction amount of each cylinder is corrected in accordance with the ratio between the sum of the absolute values of the rotation fluctuation differences of the cylinders and the rotation fluctuation difference of the cylinder, respectively. If the allowable upper limit is exceeded, the injection correction amount of the relevant cylinder is reduced by the set amount, and the injection correction amount of the other cylinder is distributed and increased by the reduced amount. If it is lower than the lower limit value, the injection correction amount of the relevant cylinder is increased by a set amount, and the injection correction amount of the other cylinder is set by distributing and decreasing the injection correction amount by the increased amount. An injection amount calculating means for correcting the basic injection amount with an injection correction amount for each cylinder when the engine shifts to an idle state; and when the engine is in an idle state, the injection correction amount of the cylinder becomes an allowable upper limit value. If it exceeds, the ignition timing correction amount of the cylinder is advanced by a predetermined angle, and if the injection correction amount of the cylinder is lower than the allowable lower limit, the ignition timing correction amount of the cylinder is set to be delayed by a predetermined angle. Ignition timing correction amount setting means; and ignition timing calculation means for calculating a basic ignition timing based on an operating state of the engine and correcting the basic ignition timing for each cylinder with the ignition timing correction amount when the engine shifts to an idle state. Is .

In this embodiment, each of the above sensors 8, 9a, 16, 18, 19, 2
0, 22, 24, the idle switch 9b and the like constitute engine operating state detecting means, and the ignition plug 11, the ignition coil 11a, the igniter 25 and the like constitute ignition means.

(Operation) Next, idle control means by the control device 31 will be described with reference to flowcharts shown in FIGS.

This idle control is based on the average difference rotation speed calculation procedure (the seventh difference rotation speed).
FIG. 8), a fuel injection control means (FIG. 8), and an ignition timing control procedure (FIG. 9). This flowchart is executed for each cylinder at each cycle time.

: Average differential rotation speed calculation procedure: First, in step (hereinafter abbreviated as “S”) S101, vehicle speed S
And the output of the idle switch is read.
It is determined whether or not the current operation state is idle based on the vehicle speed S and idle switch output read in step (1).

Vehicle speed S = 0, idle switch ON (throttle fully closed)
In the case of, it is determined that the vehicle is idling, the process proceeds to s103, and the vehicle speed S ≠
If it is 0 or the idle switch is OFF (throttle fully closed release), it is determined that the vehicle is running and the routine is ended.

In step S103, the combustion stroke cylinder #i (i = 1, 3, 2, 4) is determined from the cam pulse output from the cam angle sensor 18. Next, in S104, a crank pulse for detecting BTDC θ2, θ3 output from the crank angle sensor 16 is determined by interruption of the cam pulse.

Then, in S105, the elapsed time t2.3 between the crank pulses for detecting the BTDC θ2, θ3 determined in S104 and the θ2,
The period f2.3 is calculated from the included angle of θ3 (θ2−θ3) (f2.3 = dt2.3 / d (θ2−θ3)).

Next, in S106, the engine speed NNEW is calculated from the cycle f2.3 calculated in S105 (NNEW ← 60 / f2.3).

Thereafter, in S107, the section in which the work in the combustion stroke cylinder #i is not performed due to the combustion is performed based on the difference between the engine speed NNEW calculated in S106 and the engine speed NOLD of the cylinder # i-1 calculated in the previous routine. The difference rotational speed ΔN # i (see FIG. 6) is calculated (ΔN # i ← NNEW−NOLD), and S108
Then, the previous engine speed NOLD stored at a predetermined address in the RAM 34 is updated with the current engine speed NNEW calculated in S106 (NOLD ← NNEW).

Then, in S109, the differential rotation speed ΔN calculated in S107
Based on #i and the previous average differential rotation speed ΔNA # i (−1) of the cylinder #i stored at a predetermined address in the RAM 34, the average differential rotation speed ΔNA which is an example of the current engine rotation fluctuation difference. weighting coefficients #i determined from the weighted average of (weighted weighted average) r (ΔNA # i ← ( (2 r -1) × ΔNA # i
(-1) + ΔN # i) / 2 ^r ) and store the data at the corresponding address in the RAM 34.

Thereafter, in S110, the previous average differential rotation speed ΔNA # i (−) of the cylinder #i stored at the predetermined address of the RAM 34 in the previous time.
1) is the current average differential rotation speed ΔNA # obtained in S109.
i is updated (ΔNA # i (−1) ← ΔNA # i), and the routine is exited.

Since NOLD is not set in the first routine, S107 is jumped, and in the first routine for each cylinder, the routine is exited with ΔNA # i (−1) ← 0.

As shown in FIG. 6, in the case of a four-cycle four-cylinder engine, the engine speed NNEW obtained based on the cycle f2.3
Is performed every 180 ° C., and therefore, the cycle f2.3 measured between the cylinders is common.

Therefore, for example, when focusing on the cylinder # 1, if the previously calculated engine speed NOLD is subtracted from the currently calculated engine speed NNEW, the differential rotation speed Δ of the cylinder # 1 can be obtained.
N # 1 is obtained. On the other hand, when looking at the cylinder # 3, the engine speed NNEW of the cylinder # 1 is set to NOLD, and the difference rotational speed Δ from the engine speed NNEW of the cylinder # 3 thereafter.
N # 3 can be obtained.

Set the engine speed of the common cylinder to N4.
When 1, N, 1.3, N3.2, and N2.4 are set, the differential rotation speed of each cylinder is as follows.

ΔN # 1 = N1.3−N4.1 ΔN # 3 = N3.2−N1.3 ΔN # 2 = N2.4−N3.2 ΔN # 4 = N4.1−N2.4 Experiments have shown that ΔN # i has a strong correlation with the indicated mean effective force Pi, that is, the combustion state of the cylinder. Therefore, as described above, by calculating the differential rotation speed ΔN # i, each cylinder #
It is possible to estimate the quality of the combustion state i (illustrated average effective pressure).

The relationship between the differential rotation speed ΔN # i and the indicated mean effective pressure will be described below.

First, when the state where the engine is rotating is expressed by an equation, I: Moment of inertia N: Engine speed Ti: Command torque Tf: Friction torque And then replace it with pressure, Pi: Indicated average effective pressure Pf: Friction loss effective pressure

According to the experiment, the crank angle for detecting the rotation speed and the crank angle width for calculating the speed were set to θ2.3 as described above,
That is, if it is set before and after the combustion stroke, 4 cycles 4
In the case of a cylinder engine, based on the difference rotational speed ΔN # i and a temporal change ΔT (equivalent to 180 ° C. A rotation) during the period,
As a result of obtaining dN / dt in the above equation (3), a very strong correlation is obtained.

In this case, the fluctuation of ΔT (180 ° C. A) is negligible, and if the friction loss effective pressure Pf is considered to be constant, from the above equation (3), ΔN = K × Pi−PF (4) K , PF: The constant holds.

Accordingly, the indicated average effective pressure Pi, that is, the combustion state can be estimated for each cylinder by obtaining the differential rotation speed ΔN of each cylinder.

If the differential rotation speed ΔN # i of each cylinder #i approaches “0” individually, the combustion state of each cylinder can be made uniform.

On the other hand, in the above formula (3), assuming that the friction average effective pressure Pf is constant and is set to a constant C, and a proportional constant is set to K, Therefore, by determining K and C in advance, the indicated mean effective pressure Pi can be determined.

According to the equation (5), the indicated mean effective pressure Pi can be more accurately estimated from the difference rotation speed ΔN by differentiating the difference rotation speed ΔN with time.

Further, by calculating the average difference rotation speed ΔNA # i by a weighted average, it is possible to correct the measurement error of the cylinder and the variation due to the temporary rotation speed fluctuation.

: Fuel injection control procedure: Next, the fuel injection control procedure will be described with reference to FIG.

In addition, the battery 41 was removed due to vehicle repair, etc.,
Invalid injection correction pulse width Ts # i and ignition timing correction learning value LADV for each cylinder stored in backup RAM 35
The data of #i may be broken and become a meaningless value.
Therefore, at the time of initialization, in order to detect whether or not the battery 41 has been removed, a predetermined constant is usually stored in a specific address of the backup RAM 35, and whether or not the value of this constant is broken is stored in a characteristic address of the ROM 33. If the constant value stored in the backup RAM 35 is not different, the constant value is equal to the reference value, and the constant value is different. Assuming that the battery 41 has been disconnected, the invalid injection correction pulse width Ts # i for each cylinder is set to Ts / n (Ts; for example, the total value of the invalid injection pulse widths for each cylinder when the battery voltage is 14 V ΣTs, n; Number of cylinders, n = 4 for a four-cylinder engine,
The value of Ts / n is stored in the ROM 33 in advance. ), The ignition timing correction learning value LADV # i for each cylinder is initialized to 0, and the constant value stored at the specific address in the backup RAM 35 is reset with the reference value. If the values of the constants in the backup RAM 35 have not changed at the next startup, the initialization of the invalid injection correction pulse width Ts # i and the ignition timing learning value LADV # i is not performed.

Normally, the following processing is repeatedly executed according to the control program.

First, in S201, the engine operation state parameters are read from the output signals of the sensors 8, 9a, 16, 18, 19, 20, 22, and 24 and the idle switch 9b, and in S202, the fuel injection is performed based on the crank pulse and the cam pulse. The corresponding cylinder #i is determined.

Next, in S203, the idle switch 9b, the vehicle speed sensor 24
It is determined from the output signal of whether the current operation state is idle.

Idle switch ON (throttle fully closed) at vehicle speed S = 0
In the case of, it is determined that the vehicle is idling, the process proceeds to S204, and the vehicle speed S ≠ 0,
Alternatively, when the idle switch is OFF, it is determined that the idle is released, and the process proceeds to S205.

In S205, the fuel injection pulse width Ti of the cylinder #i corresponding to the fuel injection is obtained by the following equation as in the conventional case, and the flow jumps to S224.

Ti = Tp × α × COEF + Ts Tp = K · Q / N Tp: Basic fuel injection pulse width (basic injection amount) α: Air-fuel ratio feedback correction coefficient COEF: Various increase correction coefficients Ts: Set based on battery terminal voltage VB Voltage correction pulse width for correcting the invalid injection pulse width that has been applied Q: Intake air amount N: Engine speed K: Reciprocal of a constant based on stoichiometric air-fuel ratio, injector injection characteristics, number of cylinders, etc. In S204, the invalid injection correction pulse width Ts # i as the injection correction amount of the corresponding cylinder #i stored at the predetermined address of the backup RAM 35 is read, and in S206, the invalid injection correction pulse width Ts # i is set in advance. It compares with the set invalid injection correction pulse width allowable upper limit value TsLIMH. If TsLIMH> Ts # i, it is determined that there is room for correction on the upper limit side of the invalid injection correction pulse width Ts # i, and the process proceeds to S207. Invalid when TsLIMH ≦ Ts # i And elevation correction pulse width Ts # i is determined to have reached the upper limit value
Proceed to S208.

In S206, when it is determined that TsLIMH> Ts # i and the process proceeds to S207,
The invalid injection correction pulse width Ts # i is compared with a preset invalid lower limit value TsLIML of the invalid injection correction pulse, and TsLIML <
If Ts # i, that is, TsLIML <Ts # i <TsLIMH, it is determined that the invalid injection correction pulse width Ts # i of the cylinder #i is within the allowable limit, and the process proceeds to S209, where TsLIML ≧
In the case of Ts # i, it is determined that the invalid injection correction pulse width Ts # i has reached the lower limit, and the process proceeds to S210.

Note that the invalid injection correction pulse width allowable limit value TsLIMH, T
When the basic fuel injection amount Tp is corrected by various increase corrections such as an increase in warm-up and an increase in the air conditioner, the air-fuel ratio of the cylinder #i is over-rich even if the invalid injection correction pulse width Ts # i is added. It is set in advance by experiments or the like as a limit value that does not become over-lean or over-lean.

In S207, it is determined that TsLIML <Ts # i, and S209
Then, the program reads the average differential rotation speed ΔNA # i of the cylinder #i calculated by the average differential rotation speed calculation program, and proceeds to S
At 211, the average difference rotation speed ΔNA # i is compared with a preset average difference rotation speed allowable upper limit ΔNu, and ΔNA # i ≦
In the case of ΔNu, it is determined that the average differential rotation speed ΔNA # i is lower than the average differential rotation speed allowable upper limit value, and the process proceeds to S212, where the average differential rotation speed ΔNA # i and the preset average differential rotation speed allowable lower limit value are set. If ΔNA # i ≧ ΔNL, the average differential rotation speed ΔNA # i of the cylinder #i is within the allowable value, that is, the ideal combustion state (Δ
Nu ≧ ΔNA # i ≧ ΔNL), and the process proceeds to S223.

On the other hand, in S206, it is determined that the invalid injection correction pulse width Ts # i of the cylinder #i has reached the upper limit TsLIMH (TsLIMH ≦ Ts # i).
The ignition timing correction learning value LADV # i as the ignition timing correction amount for the cylinder #i stored at the predetermined address of No. 5 is advanced by a set crank angle C (for example, C = 1 ° C. A). Update (LADV # i ← LADV # i-C).

Next, at S213, the ignition timing correction learning value LADV # i set at S208 is compared with a preset advance limit correction value LmtADV. If LADV # i> LmtADV, the ignition timing correction learning value LADV # i is obtained. It is determined that has not yet reached the advance limit correction value LmtADV, and the routine proceeds to S215.

On the other hand, if LADV # i ≦ LmtADV, it is determined that the ignition timing correction learning value LADV # i has reached the advance limit, and the ignition timing correction learning value LADV stored at the predetermined address of the backup RAM 35 in S214. #I is the lead angle limit correction value Lm
Update with tADV (LADV # i ← LmtADV), and proceed to S215.

In S207, the invalid injection correction pulse width of the cylinder #i is determined.
Ts # i is determined to be equal to or less than the lower limit value TsLIML (TsLIML ≧ Ts #
i), proceeding to S210, the ignition timing correction learning value of the cylinder #i stored at a predetermined address in the backup RAM 35
Change LADV # i to the set crank angle C (for example, C = 1 ° C. A)
(LADV # i ← LADV # i +
C).

Next, in S216, the ignition timing correction learning value LADV # i set in S210 is compared with the preset retard limit correction value LmtRTD, and if LADV # i <LmtRTD, the ignition timing correction learning value
It is determined that LADV # i has not yet reached the retard limit correction value LmtRTD, and the process proceeds to S218.

On the other hand, if LADV # i ≧ LmtRTD, it is determined that the ignition timing correction learning value LADV # i has reached the retard limit, and S217
Then, the ignition timing correction learning value LADV # i stored at the predetermined address of the backup RAM 35 is updated with the retard limit correction value LmtRTD (LADV # i ← LmtRTD), and the routine proceeds to S218.

Note that the advance angle limit correction value LmtADV and the retard angle limit correction value LmtRTD have been obtained in advance through experiments or the like within a range that does not cause misfire when the ignition timing θIG is set.

Then, when the process proceeds from S213 or S214 to S215, the invalid injection correction pulse width Ts # i of the cylinder #i stored in the backup RAM 35 is updated with a value obtained by subtracting the pulse width ΔTs set in advance. (Ts # i ← Ts
# I−ΔTs) and the invalid injection correction pulse width Ts # (i + 1, i + 2, i + 3) of the other cylinders # (i + 1, i + 2, i + 3) are equally distributed values ΔTs / (n−1) (n) of the pulse width ΔTs. : Updated with the value obtained by adding the number of cylinders, n = 4 in the embodiment (Ts #)
(I + 1, i + 2, i + 3) ← Ts # (i + 1, i + 2, i + 3) + Δ
Ts / (n-1)), and the process proceeds to S222.

On the other hand, when proceeding from S216 or S217 to S218,
The cylinder #i stored in the backup RAM 35
(Ts # i ← Ts # i) with the value obtained by adding the invalid injection correction pulse width Ts # i by the preset pulse width Δts.
+ ΔTs), the value obtained by subtracting the invalid injection correction pulse width Ts # (i + 1, i + 2, i + 3) of the other cylinder # (i + 1, i + 2, i + 3) by the equal distribution value ΔTs / (n−1) of the pulse width ΔTs. (Ts # (i + 1, i + 2, i + 3) ← Ts # (i + 1, i + 2, i
+3) -ΔTs / (n−1)), and the process proceeds to S222.

The pulse width ΔTs is set in consideration of an output fluctuation amount when the ignition timing is corrected at the set crank angle C, a response characteristic due to fuel injection correction, and the like.

Further, the ignition timing control has a larger correction effect than the fuel injection control, and a slight correction of the ignition timing greatly changes the volumetric efficiency. On the other hand, even if the fuel injection pulse width is slightly corrected, a correction effect as large as the above-described ignition timing correction cannot be obtained. Therefore, by correcting the fuel loss when the ignition timing is advanced and correcting the fuel increase when the ignition timing is corrected, the variation in combustion between the cylinders #i can be more finely controlled. Not only can the ignition timing correction learning value LATDV # i be set even if the invalid injection correction pulse width Ts # i reaches the permissible limit value TsLIMH, TsLIML as a result of the fuel increase or decrease correction, for example. The invalid injection correction pulse width Ts corresponding to the increase or decrease in the volumetric efficiency after the retard correction or the retard correction at the set crank angle C
By reducing or increasing #i, the invalid injection correction pulse width Ts # i falls within the allowable limit value (TsLIMH> Ts # i>
TsLIML), the fuel correction can be restarted, and the controllability of the idle speed becomes wider.

Further, in S215 or S218, the other cylinder # (i + 1, i + 2, i + 3) is reduced or increased by the set value ΔTs of the invalid injection correction pulse width Ts # i of the cylinder #i.
The invalid injection correction pulse width Ts # (i + 1, i + 2, i + 3) of Δ
Since the equal distribution correction is performed at Ts / (n-1), the total air-fuel ratio does not change and the air-fuel ratio controllability is good.

On the other hand, if it is determined in S211 that the combustion state of the cylinder #i is too good (ΔNA # i> ΔNu), or if it is determined in S212 that the combustion state of the cylinder #i is too bad, the process proceeds to S219. Other cylinder # stored at a predetermined address in RAM 34
(I + 1, i + 2, i + 3) average difference rotational speed ΔNA # (i
+1, i + 2, i + 3), and in S220, the average differential overspeed ΔNA # i of the cylinder #i read in S209, and the other cylinder # (i + 1, i + 2, i +) read in S219.
3) The sum 絶 対 | ΔNA | of the absolute values of the average differential rotation speed ΔNA # (i + 1, i + 2, i + 3) is obtained (Σ | ΔNA | ← | Δ
NA # i | + | ΔNA # (i + 1) | + | ΔNA # (i +
2) | + | ΔNA # (i + 3) |).

Then, in S221, the invalid injection correction pulse width Ts # (i, i + 1, i + 2, i + 3) of each cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address of the backup RAM 35 is set to a predetermined pulse. The width ΔTs is set to the average difference rotation speed ΔNA #
Sum of absolute values of (i, i + 1, i + 2, i + 3) ３ | ΔNA | and average difference rotational speed ΔNA # of each cylinder # (i, i + 1, i + 2, i + 3)
(I, i + 1, i + 2, i + 3) ratio ΔNA # (i, i + 1, i + 2, i
+3) / Σ | ΔNA | is updated with the value distributed according to (Ts #
(I, i + 1, i + 2, i + 3) ← Ts # (i, i + 1, i + 2, i + 3) −
ΔTs × ΔNA # (i, i + 1, i + 2, i + 3) / Σ | ΔNA
|), Proceed to S222.

For example, when the vector of the average differential rotation speed ΔNA # i of the cylinder #i is on the minus side, the distribution value ΔTs × ΔNA
The vector of # i / Σ | ΔNA | becomes negative. Therefore, subtracting the distribution value ΔTs × ΔNA # i / Σ | ΔNA | from the invalid injection correction pulse width Ts # i of the cylinder #i results in In other words, the invalid injection correction pulse width Ts # i is increased and corrected.

When the vector of the average differential rotation speed ΔNA # i of the cylinder #i is on the plus side, the distribution value ΔTs × ΔNA
Since # i / Σ | ΔNA | is directly subtracted from the invalid injection correction pulse width Ts # i, the invalid injection correction pulse width
Ts # i is reduced.

Therefore, whether the combustion state of a certain cylinder #i is too good,
Alternatively, if it is too bad, the invalid injection correction pulse width of all cylinders is corrected at a ratio corresponding to the average difference rotational speed of each cylinder.

When the process proceeds from S221 to S222, each cylinder # (i, i + 1, i + 2, i +
3) Clear all of the average differential rotation speeds ΔNA # (i, i + 1, i + 2, i + 3) to obtain ΔNA # (i, i + 1, i + 2, i + 3 ← 0), S2
Proceed to 23. Since the average difference rotation speed ΔNA # i is obtained by a weighted average, the invalid injection correction pulse width Ts #
i, or even if the fuel injection or ignition timing of the cylinder #i is properly corrected with the ignition timing correction learning value LADV # i, the average differential rotation speed ΔNA # i is immediately allowed (ΔNu ≧ ΔNA).
# I ≧ ΔNL), and all cylinders # (i, i + 1,
Average difference rotational speed ΔNA # (i, i + 1, i + 2, i) of (i + 2, i + 3)
If +3) is not cleared, erroneous determination may occur in the next and subsequent operation cycles.

Then, when proceeding from S212 or S222 to S223, the basic fuel injection pulse width (basic injection amount) Tp set based on the engine operating state parameter read in S201,
Air-fuel ratio feedback correction coefficient α, various increase correction coefficients
COEF and read out in S204 or S215, S
Based on the invalid injection correction pulse width Ts # i set in S218 and S221, the fuel injection pulse width Ti of the cylinder #i is set by the following equation.

Ti ← Tp × α × COEF + Ts # i Then, in S224, the drive signal corresponding to the fuel injection pulse width Ti set in S223 is used to inject the injector of the corresponding cylinder #i.
Output to 10 at a predetermined timing, and exit the routine.

Thus, the invalid injection correction pulse width Ts # of each cylinder #i
When the average differential rotation speed ΔNA # i exceeds the allowable limit value in a state where i is within a predetermined allowable range (TsLIMH> Ts # i> TsLIML) (ΔNA # i> ΔNu or ΔNA # i <Δ)
NL), the invalid injection correction pulse widths Ts # i to Ts # (i + 3) of all the cylinders #i to # (i + 3) are distributed according to the ratio of the average differential rotation speeds ΔNA # i to ΔNA # (i + 3) of each cylinder. Pulse width ΔTs × ΔNA # i / Σ | ΔNA | ΔTs × ΔA # (i
+3) / Σ | ΔNA |, so-called differential control type correction is adopted, so that the convergence of the average differential rotational speed of each cylinder is improved without changing the total air-fuel ratio, and the idle stability is improved. I do.

: Ignition timing control procedure: The ignition timing control procedure will be described with reference to FIG.

First, in step S301, a crank pulse and a cam pulse are read, and in step S302, cylinder determination is performed based on the crank pulse and the cam pulse read in step S301.

Next, in S303, the output is output from the crank angle sensor 16.
The crank pulse for detecting BTDC θ1 and θ2 is determined from the interruption of the cam pulse.

Then, in S304, the elapsed time t1.2 between the crank pulses for detecting the BTDC θ1, θ2 determined in S303 and the θ1,
The period f1.2 is calculated from the included angle of θ2 (θ1−θ2) (f1.2 ← dt1.2 / d (θ1−θ2)).

Next, in S305, the engine speed N1.2 is calculated based on the cycle f1.2 calculated in S304 (N1.2 ← 60 / f1,2).

Thereafter, in S306, the intake air amount Q is calculated based on the output signal of the intake air amount sensor 8, and in S307, the intake air amount Q calculated in S306 and the engine speed calculated in S305 are calculated.
Based on N1.2, engine load (basic fuel injection pulse width)
Calculate Tp (Tp ← K × Q / N1.2 K: constant).

Then, in S308, the engine load Tp calculated in S307
The basic ignition timing θBASE is set based on the basic ignition timing map MPθBASE (see FIG. 5) using the engine speed N1.2 calculated in S305 as a parameter.

In S309, knock control value θNK is set according to the output signal of knock sensor 19.

Then, in S310, the output value of the vehicle speed sensor 24 and the output of the idle switch 9b are read, and in S311, if the vehicle speed S ≠ 0 or the idle switch is OFF (throttle fully closed release), it is determined that the vehicle is in the idle release state. Proceed to S312,
If the vehicle speed S = 0 and the idle switch is ON (throttle is fully closed), it is determined that the vehicle is in the idle state, and the process proceeds to S313.

In S312, the basic ignition timing θBASE set in S308 is corrected by the knock control value θNK set in S309 to calculate the ignition timing θIG (θIG ← θBASE + θNK).

On the other hand, when it is determined that the engine is idling and the process proceeds to S313, the ignition timing correction learning value LADV # i of the corresponding cylinder #i set in the above-described fuel injection control program is read out.
Is corrected by the knock control value θNK set in S309 and the ignition timing correction learning value LADV # i read out in S313 to obtain the ignition timing θIG
Is calculated (θIG ← θBASE + θNK + LADV # i).

Then, in S315, an ignition time ADV is set from the ignition timing θIG calculated in S312 or S314 and the cycle f1.2 calculated in S304 (ADV ← θIG × f1.2).

Next, in S316, the ignition time ADV set in S315 is set as a timer, and in S317, time measurement is started with the θ2 pulse as a trigger. In S318, when the ignition time is reached, the corresponding cylinder #i
The ignition signal is output to the igniter 25, and the routine is ended.

Next, the invalid injection correction pulse width Ts according to the above program
#I and the setting state of the ignition timing correction learning value LADV # i
This will be described with reference to the time charts of FIGS.

For example, when the combustion state of the cylinder # 1 is poor and the average differential rotation speed ΔNA # i becomes lower than the allowable lower limit value ΔNL of the average differential rotation speed, the average differential rotation of all the cylinders # (1,3,2,4) at this time. The pulse width ΔTs preset according to the ratio between the sum of the absolute values of the speeds ΔNA # (1,3,2,4) Σ | ΔNA | and each average difference rotational speed ΔNA # (1,3,2,4) Is distributed to each cylinder # (1,3,2,4), (ΔTs × ΔNA # (1,3,2,4) / Σ | ΔNA |), and each of the cylinders # (1, Correct the invalid injection correction pulse width Ts # (1,3,2,4) of (3,2,4) (Ts # (1,3,2,
4) ← Ts # (1,3,2,4) −ΔTs × ΔNA # (1,3,2,4) / Σ
| ΔNA |).

Then, as shown in the figure, the invalid injection correction pulse Ts # i of the cylinder # 1 having a bad combustion state is increased and corrected, and the invalid injection correction pulse width Ts of the cylinders # 3 and # 2 having a relatively good combustion state is corrected.
# 3 and Ts # 2 are reduced by the distribution values corresponding to the average difference rotational speeds ΔNA # 3 and ΔNA # 2. Further, in the cylinder # 4 in which combustion is stable, the invalid injection correction pulse width Ts # 4 is not substantially corrected because the average difference rotational speed ΔNA # 4 is zero.

As a result, the average differential rotation speed Δ of each cylinder # (1,3,2,4)
The convergence of NA # (1,3,2,4) is improved.

Then, the average differential rotation speed ΔNA of each cylinder # (1,3,2,4)
#Reset (1,3,2,4) (elapsed time t1, t2, t3, t
Four).

After that, the invalid injection correction pulse width Ts # 1 of the cylinder # 1
Reaches the allowable upper limit TsLIMH of the invalid injection correction pulse width, the set pulse width ΔTs is subtracted from the invalid injection correction pulse width Ts # 1 (Ts # 1 ← Ts # 1-ΔTs), and the relevant cylinder # 1
The fuel injection pulse width Ti of the other cylinders #
ΔTs for invalid injection correction pulse width Ts # (3,2,4) of (3,2,4)
/ 3 (Ts # (3,2,4) Ts # (3,2,4) + ΔTs / 3)
Further, the ignition timing correction learning value LADV # 1 of the cylinder # 1 is advanced by the set crank angle C (LADV # 1 ← LADV # 1).
-C). Next, the average difference rotational speed ΔNA # (1,3,2,4) of each cylinder # (1,3,2,4) is reset (elapsed time t5).

 As a result, the same effect as described above can be obtained.

Further, as shown in the elapsed times t6 and t7, when the average differential rotation speed ΔNA # 3 becomes higher than the average differential rotation allowable upper limit ΔNu because the combustion state of the cylinder # 3 is too good, all the cylinders # (1,
3,2,4) and the sum of absolute values 絶 対 | ΔNA | of the average differential rotational speeds ΔNA # (1,3,2,4) and the average differential rotational speeds ΔNA of the cylinders # (1,3,2,4). Correcting each cylinder # (1,3,2,4) with a correction value distributed according to the ratio to # (1,3,2,4) (Ts # (1,3,2,4)
← Ts # (1,3,2,4) −ΔTs × ΔNA # (1,3,2,4) / Σ |
ΔNA |).

Then, the invalid injection correction pulse width Ts # 3 of the cylinder # 3 in the good combustion state is reduced and the invalid injection correction pulse width Ts # 3 of the other cylinder # (2,4,1) whose average differential rotation speed vector is on the minus side is corrected. The correction pulse width Ts # i (2,4,1) is increased by a correction value corresponding to the average difference rotational speed ΔNA # (2,4,1). Then, the average difference rotational speed ΔNA # (1,3,2,4) of each cylinder # (1,3,2,4) is reset.

Thereafter, the invalid injection correction pulse width Ts # 3 of the cylinder # 3
Reaches the invalid lower limit TsLIML of the invalid injection correction pulse width, the pulse width Δ is set to the invalid injection correction pulse width Ts # 3.
Ts is added (Ts # 3 ← Ts # 3 + ΔTs), the fuel injection pulse width Ti of the cylinder # 3 is increased, and the other cylinders #
From the invalid injection correction pulse width Ts # (2,4,1) of (2,4,1)
Subtract Ts / 3 (Ts # (2,4,1) ← Ts # (2,4,1) −ΔTs /
3) Further, the ignition timing correction learning value LADV of the cylinder # 3 concerned
# 3 is retarded at the set crank angle C (LADV # 3 ← LA
DV # 3 + C). Next, the average difference rotational speed ΔNA # (1,3,2,4) of each cylinder # (1,3,2,4) is reset (elapsed time).
t8).

(Second Embodiment) FIG. 12 is a flowchart showing a cylinder-by-cylinder fuel injection control procedure according to a second embodiment of the present invention. Steps similar to those in FIG. 8 of the first embodiment are denoted by the same reference numerals, and description thereof is omitted.

In this embodiment, first, the fuel injection pulse width Ti is obtained,
During idling, the fuel injection pulse width Ti is corrected by a correction pulse width Tc # i as an injection correction amount to perform idle control. Note that the correction pulse width Tc # i is stored at a predetermined address of the backup RAM 35 corresponding to each cylinder, and the initial set value is 0 when data is broken. The procedure for calculating the average differential rotation speed ΔNA # i is the same as that in the first embodiment (FIG. 7), and thus the description is omitted.

First, when the process proceeds from S201 and S202 to S401, the fuel injection parameter width Ti is obtained based on the engine operation state parameter detected in S201 (Ti ← Tp × α × COEF + Ts).

Next, in S203, it is determined whether or not the current operation state is idle, and if it is in the idle release state, the process jumps to S224. On the other hand, if it is determined that the cylinder is in the idle state, the process proceeds to S402, and the corresponding cylinder # stored at a predetermined address in the backup RAM 35
i, the correction pulse width Tc # i is read, and in step S403, the correction pulse width Tc # i and the preset correction pulse width allowable upper limit value TcLI are set.
MH, and when TcLIMH> Tc # i, the correction pulse width Tc
It is determined that there is room for correction on the upper limit side of #i, and the process proceeds to S404. If TcLIMH ≦ Tc # i, the correction pulse width Tc # i
Is determined to have reached the upper limit, and the flow proceeds to S208.

In S403, when it is determined that TcLIMH> Tc # i and the process proceeds to S404,
The correction pulse width Tc # i is compared with a preset correction pulse width allowable lower limit value TcLIML. If TcLIML <Tc # i, that is, TcLIML <Tc # i <TcLIMH, the correction pulse width Tc of the cylinder #i is used. It is determined that #i is within the allowable limit, and the process proceeds to S209. If TcLIML ≧ Tc # i, it is determined that the correction pulse width Tc # i has reached the lower limit, and the process proceeds to S210.

It should be noted that the corrected pulse width allowable limit values TcLIMH and TcLIML can be obtained by adding this value to the fuel injection pulse width Ti even if the cylinder #i
The air-fuel ratio is set as a limit value obtained from experiments or the like in advance as a limit value at which the air-fuel ratio does not become overrich or overlean.

If it is determined in step S404 that TcLIML <Tc # i and the process proceeds to step S209, the average differential rotation speed ΔNA # i of the cylinder #i is read out, and in step S211 the average differential rotation speed ΔNA # i and the preset average differential rotation By comparing with the allowable speed upper limit value ΔNu, ΔNA #
If i ≦ ΔNu, the process proceeds to S212, and compares with the average difference rotational speed allowable lower limit value ΔNL, and ΔNA # i ≧ ΔNL, that is, ΔN
If u ≧ ΔNA # i ≧ ΔNL, the process proceeds to S408.

Further, when the process proceeds from S403 to S405 via S208, S213, and S214, a value obtained by subtracting the correction pulse width Tc # i of the cylinder #i stored at a predetermined address of the backup RAM 35 by the preset pulse width ΔTc. Update with (Tc #
i ← Tc # i−ΔTc), other cylinders # (i + 1, i + 2, i + 3)
Is updated by the value obtained by adding the corrected pulse width Tc # (i + 1, i + 2, i + 3) by the equal distribution value ΔTc / (n−1) (n is the total number of cylinders, n = 4 in the embodiment) of the pulse width ΔTc. (Tc # (i
+1, i + 2, i + 3) ← Tc # (i + 1, i + 2, i + 3) + ΔTc /
(N-1)), and proceed to S222.

On the other hand, when the process proceeds from S404 through S210, S216, S217 to S406, the value obtained by adding the correction pulse width Tc # 1 of the cylinder #i stored at the predetermined address of the backup RAM 35 by the preset pulse width ΔTc Update with (Tc #
i ← Tc # i + ΔTc), other cylinders # (i + 1, i + 2, i + 3)
Is updated with a value obtained by subtracting the corrected pulse width Tc # (i + 1, i + 2, i + 3) by the equal distribution value ΔTc / (n−1) of the pulse width ΔTc (Tc # (i + 1, i + 2, i + 3) ← Tc # (I + 1, i +
(2, i + 3) −ΔTc / (n−1)), and the process proceeds to S222.

The pulse width ΔTc is set in consideration of an output fluctuation amount when the ignition timing is corrected at the set crank angle C, a response characteristic due to fuel injection correction, and the like.

Further, in step S211 it is determined that ΔNA # i> ΔNu, or in step S212 it is determined that ΔNA # i <ΔNL, and 219, S220
And the process proceeds to S407, the cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address in the backup RAM 35
The correction pulse width Tc # (i, i + 1, i + 2, i + 3) is calculated by the sum of absolute values of the preset pulse width ΔTc and the average difference rotational speed ΔNA # (i, i + 1, i + 2, i + 3) set in S220. | Δ
NA | and the ratio ΔNA # (i, i) of the average differential rotation speed ΔNA # (i, i + 1, i + 2, i + 3) of each cylinder # (i, i + 1, i + 2, i + 3).
+1, i + 2, i + 3) / Σ | ΔNA |, and updates with a value distributed according to (Tc # (i, i + 1, i + 2, i + 3) ← Tc # (i, i + 1, i
+ 2, i + 3) −ΔTc × ΔNA # (i, i + 1, i + 2, i + 3) /
Σ | ΔNA |), and proceeds to S222.

For example, when the vector of the average differential rotation speed ΔNA # i of the cylinder #i is on the minus side, the distribution value ΔTc × Δ
The vector of NA # i / Σ | ΔNA | is negative, and therefore, subtracting the distribution value ΔTc × ΔNA # i / Σ | ΔNA | from the correction pulse width Tc # i of the cylinder #i results. Then, the correction pulse width Tc # i is increased and corrected.

When the vector of the average differential rotation speed ΔNA # i of the cylinder #i is on the plus side, the distribution value ΔTc × ΔNA
Since # i / Σ | ΔNA | is directly subtracted from the sanity correction pulse width Tc # i, the correction pulse width Tc # i is reduced.

As a result, the same functions and effects as those of the first embodiment can be obtained.

Thereafter, when the process proceeds from S405, S406, or S407 to S222, the average differential rotation speed ΔNA # (i, i, i) of each cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address in the RAM 34.
Clear all i + 1, i + 2, i + 3) (ΔNA # (i, i +
1, i + 2, i + 3) ← 0), and the process proceeds to S408.

Then, when proceeding from S222 or S212 to S408, the fuel injection pulse width Ti set in S401 is changed to S402.
Or the correction is performed using the correction pulse width Tc # i set in S405, S406, and S407 (Ti ← Ti + Tc # i), and in S224, the drive signal corresponding to the fuel injection pulse width Ti is output to the cylinder #i. Output to the injector 10 at a predetermined timing,
Out of the routine.

The procedure for controlling the ignition timing in this embodiment is the same as that in the first embodiment (FIG. 9) described above, and a description thereof will be omitted.

(Third Embodiment) FIGS. 13 to 15 show a third embodiment of the present invention, FIG. 13 is a flowchart showing a cylinder-by-cylinder average differential angular acceleration calculation procedure, and FIG. 14 is a cylinder-by-cylinder fuel injection control procedure. FIG. 15 is a time chart of the crank pulse, the cam pulse, the engine speed and the angular acceleration.

In this embodiment, the rotational fluctuation difference of each cylinder #i is obtained from the average differential angular acceleration Δ (dN / dt) A # i instead of the average differential rotational speed ΔNA # i shown in the first and second embodiments. It is what we asked for.

: Average difference angular acceleration calculation procedure: Since S101 to S105 are common to the first embodiment (FIG. 7),
Only the flowcharts after S106 (portions surrounded by broken lines in FIG. 7) will be described based on FIG.

In S106, when the engine speed NNEW is calculated based on the cycle f2,3 set in S105, the process proceeds to S501, in which the engine speed NNEW is time-differentiated to calculate an angular acceleration (dN / dt) NEW.

Thereafter, in S502, the combustion stroke cylinder #i is determined from the difference between the engine angular acceleration (dN / dt) NEW calculated in S501 and the engine angular acceleration (dN / dt) OLD calculated in the previous routine.
The section that does not work due to the combustion of the section (section θ2 in FIG. 15)
−θ3) to calculate a differential angular acceleration Δ (dN / dt) #i (Δ (d
N / dt) # i ← (dN / dt) NEW− (dN / dt) OLD), at S503,
The previous engine angular acceleration (dN / dt) OLD stored at a predetermined address in the RAM 34 is updated with the current engine angular acceleration (dN / ds) NEW calculated in S501 ((dN / dt) OLD ←
(DN / dt) NEW).

Then, in S504, the difference angular acceleration Δ calculated in S502 is obtained.
Based on (dN / dt) #i and the previously calculated average differential angular acceleration Δ (dN / dt) A # i (−1) of the cylinder #i, the current average differential angular acceleration Δ (dN / dt) Ai Is obtained from the weighted average of the weighting coefficient r shown in the following equation, and stored in the corresponding address of the RAM 34.

Δ (dN / dt) A # i ← ((2 r -1) × Δ (dN / dt) A # i (-1) + Δ (dN / dt) #i) / 2 r where mean difference angle initial The acceleration Δ (dN / dt) A # i is “0”.

Thereafter, in S505, the average differential angular acceleration Δ (dN / dt) A # i of the cylinder #i calculated in S504 is the average differential angular acceleration Δ of the cylinder #i stored at a predetermined address in the RAM 34.
(DN−dt) A # i (−1) is updated (Δ (dN / dt) A
#I (-1) ← Δ (dN / dt) A # i).

: Fuel injection control procedure: As described above, in this embodiment, the average difference angular acceleration Δ (dN / d
t) Since the rotation fluctuation of each cylinder #i is determined based on A # i, the first embodiment (FIG. 8) and the second embodiment (12th embodiment)
Only the flowchart of the part surrounded by the broken line in FIG.

That is, when the process proceeds from S207 (or S404) to S601, the average angular acceleration Δ (dN / dt) A # i of the cylinder #i is read, and in S602, the average angular acceleration Δ (dN / dt) A # i.
And the predetermined upper limit of the average angular acceleration allowable Δ (dN / dt) u
When Δ (dN / dt) A # i ≦ Δ (dN / dt) u, the process proceeds to S603, and the average differential angular acceleration Δ (dN / dt) A # i
And the average lower limit of angular acceleration Δ (dN / dt) L,
Δ (dN / dt) A # i ≧ Δ (dN / dt) L, that is, Δ (dN
/ dt) u ≧ Δ (dN / dt) A # i ≧ Δ (dN / dt) L
Proceed to 223 (or S408).

On the other hand, in S602, it is determined that Δ (dN / dt) A # i> Δ (dN / dt) u, or in S603, Δ (dN / dt) A # i <Δ (dN /
If it is determined that dt) L, the process proceeds to S605, and the other cylinders # (i + 1, i + 2, i +
3) Average difference angular acceleration Δ (dN / dt) A # (i + 1, i + 2,
i + 3), and in S606, the average differential angular velocity Δ (dN / dt) A # i of the cylinder #i read in S601 and the other cylinders # (i + 1, i + 2, i + 3) read in S605. The sum 絶 対 | Δ (dN / dt) A | of the absolute values of the average difference angular acceleration Δ (dN / dt) A # (i + 1, i + 2, i + 3) is obtained (Σ | Δ (dN /
dt) A | ← | Δ (dN / dt) A # i | + | Δ (dN / dt) A # (i
+1) | + | Δ (dN / dt) A # (i + 2) | + | Δ (dN /
dt) A # (i + 3) |).

Then, in S607, the sum 絶 対 of the absolute values of the invalid injection correction pulse widths Ts # (i, i + 1, i + 2, i + 3) of the cylinders # (i, i + 1, i + 2, i + 3) stored at the predetermined addresses in the backup RAM 35. | Δ (dN / dt) A | and each cylinder (i, i + 1, i + 2, i +
3) Average difference angular acceleration Δ (dN / dt) A # (i, i + 1, i + 2, i
+3) and Δ (dN / dt) A # (i, i + 1, i + 2, i + 3)
/ Σ | Δ (dN / dt) A | is updated with the value distributed according to (Ts
# (I, i + 1, i + 2, i + 3) ← Ts # (i, i + 1, i + 2, i + 3)
−ΔTs × Δ (dN / dt) A # (i, i + 1, i + 2, i + 3) / Σ |
Δ (dN / dt) A |), and proceed to 0S64.

Alternatively, in the case of the correction pulse width Tc corresponding to the second embodiment, the correction pulse width Tc # (i, i + 1) of the cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address of the backup RAM 35. , i + 2, i + 3), the preset pulse width ΔTc is the sum of the absolute values of the average difference angular acceleration Σ | Δ (dN / d
t) A | and the average differential angular acceleration Δ (dN / dt) A # (i, i
+1 and i + 2, i + 3) and update with the value distributed according to the ratio of (Tc # (i, i + 1, i + 2, i + 3) ← Tc # (i, i + 1, i + 2, i +
3) -ΔTc × Δ (dN / dt) A # (i, i + 1, i + 2, i + 3) /
Σ | Δ (dN / dt) A |), and proceeds to S604.

Then, the above S215 (or S405), S218 (or
When the process proceeds from S607 to S604 or S604, each cylinder # (i, i + 1, i + 2, i +) stored at a predetermined address in the RAM 34
3) Average difference angular acceleration Δ (dN / dt) A # (i, i + 1, i + 2, i
+ (3) is cleared and ((ΔdN / dt) A # (i, i + 1, i
+2, i + 3) ← 0), and the process proceeds to S223 (or S408).

FIG. 15 shows the correlation between the engine speed and the engine angular acceleration. As described above, the time element is added by using the angular acceleration obtained by differentiating the engine speed over time. Then, the accuracy is further improved.

Further, the ignition timing control is the same as that of the first embodiment, and the description is omitted.

(Fourth Embodiment) FIGS. 16 and 17 show a fourth embodiment of the present invention, and FIG. 16 is a flowchart showing a cylinder average difference cycle calculation procedure.
FIG. 17 is a flowchart showing the fuel injection control procedure for each cylinder.

In this embodiment, the rotational fluctuation difference of each cylinder #i is calculated by using the average difference rotational speed ΔNA # i shown in the first and second embodiments,
Alternatively, the average difference angular acceleration Δ (dN / d
t) Instead of A # i, it is determined from the average difference period ΔfA # i.

: Average difference cycle calculation procedure: Since S101 to S105 are common to the first embodiment (FIG. 7), description thereof will be omitted, and only a flowchart corresponding to a portion surrounded by a broken line in FIG. 7 will be described.

When the cycle f2,3 is calculated in S105 (see FIG. 7), S7
The process proceeds to step 01, where the current cycle f2,3 calculated in S105 and the cycle f2,3OLD calculated in the previous routine stored at a predetermined address in the RAM 34 are read, and the two cycles f2,3, f2,3OLD are compared. From the difference, the combustion stroke cylinder #i (i = 1,3,
2,4) Difference period Δf of the section not working by combustion
#I is calculated (Δf # i ← f2,3-f2,3OLD).

Next, in S702, the previous cycle f2,3OLD stored at a predetermined address in the RAM 34 is updated with the current cycle f2,3 calculated in step S105 (f2,3OLD ← f2,3).

Then, in S703, the difference cycle Δf # i calculated in S701
And the previous average difference cycle ΔfA # i (−1) of the cylinder #i.
, The current average difference period ΔfA # i is calculated from the weighted average of the weighting coefficient r shown in the following equation.

ΔfA # i ← ((2 ^r −1) × ΔfA # i (−1) + Δf # i) / 2
^r Note that the initial average difference period ΔfA # i is “0”.

Thereafter, in S704, the previous average difference cycle ΔfA # i (−−) of the cylinder #i stored at a predetermined address in the RAM 34 with the average difference cycle ΔfA # i of the cylinder #i calculated in S703.
1) is updated (ΔfA # i (−1) ← ΔfA # i).

: Fuel injection control procedure: In this embodiment, since the rotation fluctuation of each cylinder #i is determined based on the average difference period ΔfA # i, the first embodiment (FIG. 8) and the second embodiment (FIG. Only the flowchart of the part surrounded by the broken line in FIG. 12) is different.

That is, when proceeding from S207 (or S404) to S801,
The average difference cycle ΔfA # i of the cylinder #i is read, and in S802,
The average difference period ΔfA # i is compared with a preset average difference period allowable shortest value Δfs, and if ΔfA # i ≦ Δfs, the process proceeds to S803 and the average difference period ΔfA # i and the average difference period allowable maximum value Δ
fL, ΔfA # i ≦ ΔfL, that is, Δfs ≦ ΔfA
In the case of # i ≦ ΔfL, the value is within the allowable value, and therefore S223
(Or S408).

On the other hand, in S802, it is determined that ΔfA # i <Δfs, that is, the average difference cycle ΔfA # i of the cylinder #i is too short, and in S803, ΔfA # i> ΔfL, that is, the average difference of the cylinder #i If it is determined that the cycle ΔfA # i is too long, the process proceeds to S805, and each average difference cycle ΔfA # (i + 1, i) of the other cylinders # (i + 1, i + 2, i + 3) stored at a predetermined address in the RAM 34.
i + 2, i + 3), and in S806, the average difference cycle ΔfA # i of the cylinder #i read in S801, and the average difference cycle ΔfA of the other cylinders # (i + 1, i + 2, i + 3) read in S805. # (I + 1, i + 2, i + 3) sum of absolute values Σ | ΔfA |
(Σ | ΔfA | ← | ΔfA # i | + | ΔfA # (i +
1) | + | ΔfA # (i + 2) | + | ΔfA # (i + 3)
|).

In step S807, the invalid injection correction pulse width Ts # (i, i + 1, i + 2, i + 3) of each cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address in the backup RAM 35 is set to a predetermined pulse. The width ΔTs is a ratio ΔfA # between the sum of the absolute values of the average difference periods Σ | ΔfA | and the average difference period ΔfA # i (i, i + 1, i + 2, i + 3) of each cylinder # (i, i + 1, i + 2, i + 3).
Update with the value distributed according to (i, i + 1, i + 2, i + 3) / Σ | ΔfA | (Ts # (i, i + 1, i + 2, i + 3) ← Ts # (i, i + 1,
i + 2, i + 3) + ΔTs × ΔfA # (i, i + 1, i + 2, i + 3) /
Σ | ΔfA |), and proceeds to S804.

Alternatively, in the case of the correction pulse width Tc corresponding to the second embodiment, the correction pulse width Tc # (i, i + 1) of the cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address of the backup RAM 35. , i + 2, i + 3) according to the ratio of the preset pulse width ΔTc to the sum of the absolute values of the average difference periods Σ | ΔfA | and the average difference period ΔfA # (i, i + 1, i + 2, i + 3) of each cylinder. Update with the distributed value (Tc # (i, i + 1, i + 2, i +
3) ← Tc # (i, i + 1, i + 2, i + 3) + ΔTc × ΔfA # (i, i
+1, i + 2, i + 3) / Σ | ΔfA |), and proceeds to S804.

When proceeding from S215 (or S405), S218 (or S406), or S807 to S804, each cylinder # (i, i + 1, i + 2,
After clearing all the average difference periods ΔfA # (i, i + 1, i + 2, i + 3) of (i + 3) (ΔfA # (i, i + 1, i + 2, i + 3) ← 0), S
Proceed to 223 (or S408).

In this embodiment, when detecting fluctuations in the engine speed, the engine speed NNEW obtained by temporarily dividing the period f2,3 is used.
Is not used, and the periods f2 and f3 are used as they are, so that the calculation load on the CPU due to the division processing is reduced, and the calculation time is shortened.

(Fifth Embodiment) FIG. 18 et seq. Show a fifth embodiment of the present invention. FIG. 18 is a flowchart showing a cylinder-by-cylinder average differential angular velocity calculation procedure.
The figure is a flowchart showing the fuel injection control procedure for each cylinder.
FIG. 20 is a flowchart showing a procedure for controlling ignition timing for each cylinder.

In this embodiment, the rotational fluctuation difference of each cylinder #i is obtained from the average difference angular velocity ΔωA # i.

: Average difference angular velocity calculation procedure: Since S101 to S104 are common to the first embodiment (FIG. 7), description thereof will be omitted.

After determining the crank pulse in S104, in S901, the elapsed time t2,3 between the crank pulses for detecting the BTDC θ2, θ3 determined in S104, and the included angle (θ2-θ3) between θ2, θ3.
And the angular velocity ω2,3NEW is calculated (ω2,3NEW ← d (θ2
−θ3) / dt2,3).

Thereafter, in S902, no work is performed by the combustion of the combustion stroke cylinder #i from the difference between the engine angular speed ω2,3NEW of the current routine calculated in S901 and the engine angular speed ω2,3OLD calculated in the previous routine. Difference angular velocity of the section Δω
#I is calculated (Δω # i ← ω2,3NEW−ω2,3OLD), and S903
Then, the previous engine angular velocity ω2,3OLD stored at the predetermined address in the RAM 34 is updated with the current engine angular velocity ω2,3NEW calculated in S901 (ω2,3OLD ← ω2,3NEW).

Then, in S904, the difference angular velocity Δω # calculated in S902 above
Based on i and the previous average differential angular velocity ΔωA # i (−1) of the cylinder #i stored at a predetermined address in the RAM 34, the current average differential angular velocity ΔωA # i is weighted by a weighting coefficient r shown in the following equation. It is obtained from the average and stored in the corresponding address of the RAM.

ΔωA # i ← ((2 ^r −1) × ΔωA # i (−1) + Δω #
i) / 2 ^r It is to be noted that the average difference between the angular velocity ΔωA # i for the first time is set to "0".

Thereafter, in S905, the data stored at the predetermined address of the RAM 34 is updated with the average differential angular velocity ΔωA # i of the cylinder #i calculated in S904 (ΔωA # i (−1) ← Δ).
ωA # i).

In this embodiment, since the rotation fluctuation of each cylinder #i is determined based on the average difference angular velocity ΔωA # i, the first embodiment (FIG. 8) and the second embodiment (FIG. 8) Only the flowchart of the part surrounded by the broken line in FIG. 12) is different.

That is, when the process proceeds from S207 (or S404) to S1001, the average angular velocity ΔωA # i of the cylinder #i is read, and S1 is read.
In 002, the average angular velocity ΔωA # i is compared with a predetermined average angular velocity upper limit Δωu. If ΔωA # i ≦ Δωu, the process proceeds to S1003, where the average difference angular velocity ΔωA # i is compared with the average angular velocity lower limit ΔωL. And ΔωA # i ≧ ΔωL,
That is, in the case of Δωu ≧ ΔωA # i ≧ ΔωL, the process proceeds to S223 (or S408) because it is within the allowable value.

On the other hand, if it is determined in S1002 that ΔωA # i <Δωu, and if it is determined in S1003 that ΔωA # i <ΔωL, S10
Proceeding to 05, the average differential angular velocity ΔωA # of each of the other cylinders # (i + 1, i + 2, i + 3) stored at the predetermined address in the RAM 34
(I + 1, i + 2, i + 3) are read, and in S1006, the average differential angular velocity ΔωA # i of the cylinder #i read in S1001 and the other cylinder # (i + 1, i + 2, i +) read in S1005.
3) Find the sum 絶 対 | ΔωA | of the absolute values of the average difference angular velocity ΔωA # (i + 1, i + 2, i + 3) (Σ | ΔωA | ← | ΔωA)
#I | + | ΔωA # (i + 1) | + | ΔωA # (i + 2)
| + | ΔωA # (i + 3) |).

Then, in S1007, the invalid injection correction pulse width Ts # (i, i + 1, i + 2, i + 3) of each cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address in the backup RAM 35 is set to a predetermined pulse. The width ΔTs is the ratio Δ between the sum of the absolute values of the average difference angular velocity Σ | ΔωA | and the average difference angular velocity ΔωA # (i, i + 1, i + 2, i + 3) of each cylinder # (i, i + 1, i + 2, i + 3).
ωA # (i, i + 1, i + 2, i + 3) / Σ | ΔωA |, updated with the value distributed according to (Ts # (i, i + 1, i + 2, i + 3) ← Ts #
(I, i + 1, i + 2, i + 3) −ΔTs × ΔωA # (i, i + 1, i +
2, i + 3) / Σ | ΔωA |), and proceeds to S1004.

Alternatively, in the case of the correction pulse width Tc corresponding to the second embodiment, the correction pulse width Tc # (i, i + 1) of the cylinder # (i, i + 1, i + 2, i + 3) stored at a predetermined address of the backup RAM 35. , i + 2, i + 3) according to the ratio of the preset pulse width ΔTc to the sum of the absolute values of the average differential angular velocity Σ | ΔωA | and the average differential angular velocity ΔωA # (i, i + 1, i + 2, i + 3) of each cylinder. Update with the distributed value (Tc # (i, i + 1, i + 2,
i + 3) ← Tc # (i, i + 1, i + 2, i + 3) + ΔTc × ΔωA #
(I, i + 1, i + 2, i + 3) / Σ | ΔωA |), and proceeds to S1004.

Then, S215 (or S405), S218 (or S40)
6) Alternatively, when the process proceeds from S1007 to S1004, each cylinder # (i, i + 1, i + 2, i +
Clear all the average difference angular velocity ΔωA # (i, i + 1, i + 2, i + 3) of (3) (ΔωA # (i, i + 1, i + 2, i + 3) ←)
0), and proceed to S223 (or S408).

: Ignition timing control procedure: In the first embodiment (FIG. 9), the ignition timing is set based on the periods f1, 2; however, in this embodiment, the angular velocity ω is set.
The difference is that the ignition timing is set based on 1 and 2.

After determining the crank pulse in S303, in S1101, the elapsed time t1,2 between the crank pulses for detecting the BTDC θ1, θ2 determined in S303 and the angle between the θ1, θ2 (θ1-θ
2) to calculate the angular velocity ω1,2 (ω1,2 ← d (θ1-
θ2) / dt1,2).

Next, in S1102, the angular velocity ω1,2 calculated in S1101 above
The engine speeds N1, 2 are calculated based on the above. (N1,2 ← 60 ×
ω1,2).

In S1103, the ignition time ADV is set from the ignition timing θIG calculated in S312 or S314 and the angular velocity ω1,2 calculated in S1101 (ADV ← θIG / ω1,2).

According to this embodiment, since the combustion state estimation is performed using the angular velocities ω2,3 instead of the engine speed NNEW, the calculation time for the combustion state estimation is reduced as compared with the first embodiment.

[Effects of the Invention] As described above, according to the present invention, the rotation fluctuation at the time of idling is relatively controlled instead of individually controlled by the fuel injection amount and the ignition timing. Not only can the combustion state be controlled more precisely, but the control can be performed over a wide range without significantly over-lean or over-rich for each cylinder, thereby greatly improving controllability and obtaining a good feeling. Excellent effects are achieved.

Further, when the rotation fluctuation difference between the cylinder and the cylinder is out of the rotation fluctuation allowable value, the correction injection of each cylinder is made according to the ratio of the sum of the absolute values of the rotation fluctuation difference of each cylinder and the rotation fluctuation difference of each cylinder. Since the so-called differential control correction for correcting the amount is employed, the convergence of the rotational fluctuation difference of each cylinder is improved without changing the total air-fuel ratio, and the idle stability is improved.

[Brief description of the drawings]

FIG. 1 is a block diagram corresponding to claims showing the basic concept of the present invention, and FIGS. 2 to 11 show a first embodiment of the present invention.
3 is a front view of a crank rotor and a crank angle sensor, FIG. 4 is a front view of a cam rotor and a cam angle sensor, FIG. 5 is a conceptual view of a basic ignition timing map, FIG. Figure 6 shows the cylinder pressure fluctuation, crank pulse,
FIG. 7 is a flowchart showing a cylinder-by-cylinder average differential rotational speed calculation procedure, FIG. 8 is a flowchart showing a cylinder-by-cylinder fuel injection control procedure, and FIG. 9 is a cylinder-by-cylinder ignition timing control. Flow chart showing the procedure, FIG. 10 is a time chart showing the average differential rotational speed for each cylinder, the invalid injection pulse width for each cylinder, FIG. 11 is a time chart showing the learned ignition timing correction value for each cylinder, and FIG. 13 to 15 show a third embodiment of the present invention, FIG. 13 is a flowchart showing a cylinder-by-cylinder average differential angular acceleration calculation procedure, and FIG. Fig. 14 is a flowchart showing the fuel injection control procedure for each cylinder, Fig. 15 is a time chart of crank pulse, cam pulse, engine speed and angular acceleration, and Figs. Shows a fourth embodiment of the sixteenth
The figure is a flowchart showing the average difference cycle calculation procedure for each cylinder,
FIG. 17 is a flowchart showing a cylinder-by-cylinder fuel injection control procedure, FIG. 18 et seq. Show a fifth embodiment of the present invention, FIG. 18 is a flowchart showing a cylinder-by-cylinder average differential angular velocity calculation procedure,
FIG. 19 is a flowchart showing the fuel injection control procedure for each cylinder,
FIG. 20 is a flowchart showing a procedure for controlling ignition timing for each cylinder. M1 ... engine operating state detecting means, M2 ... crank angle detecting means, M3 ... idling determining means, M4 ... combustion state detecting means, M5 ... injection correction amount setting means, M6 ... injection amount calculating means, M7 ... fuel supply means, M8 ... ignition timing correction amount setting means, M9 ... ignition timing calculation means, M10 ... ignition means.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code FI F02P 5/15 F02P 5/15 K

Claims (1)

(57) [Claims] 1. An operating state detecting means for detecting an operating state of an engine; a crank angle detecting means for outputting a signal capable of determining a cylinder in a combustion stroke and a signal corresponding to a predetermined crank angle; And an idle determination means for determining from the engine operating state, an engine rotation fluctuation difference for each predetermined crank angle when the engine is in an idle state, and comparing the engine rotation fluctuation difference with a preset rotation fluctuation allowable limit. A combustion state detecting means for detecting a combustion state of a cylinder in a combustion stroke; and, when the rotation fluctuation difference of the cylinder is out of a rotation fluctuation allowable limit when the engine is in an idle state, an injection correction amount of each cylinder is determined. Correction is made according to the ratio between the sum of the absolute values of the rotational fluctuation differences of the cylinders and the rotational fluctuation difference of the cylinder, respectively. If the allowable upper limit is exceeded, the injection correction amount of the relevant cylinder is reduced by the set amount, and the injection correction amount of the other cylinder is distributed and increased by the reduced amount, and the injection correction amount of the relevant cylinder is lower than the allowable lower limit value. In this case, the injection correction amount of the relevant cylinder is increased by a set amount, and the injection correction amount of the other cylinder is set by distributing and decreasing the injection correction amount by the increased amount, and the basic injection amount is set based on the operating state of the engine. The fuel injection means calculates the basic injection amount for each cylinder with the injection correction amount when the engine shifts to the idle state, and the fuel supply means supplies fuel to the corresponding cylinder based on the output of the injection amount calculation means. In the idling state, if the injection correction amount of the cylinder exceeds the allowable upper limit, the ignition timing correction amount of the cylinder is advanced by a predetermined angle, and if the injection correction amount of the cylinder is lower than the allowable lower limit, Point of the cylinder An ignition timing correction amount setting means for setting the ignition timing correction amount by retarding the ignition timing by a predetermined angle; and calculating a basic ignition timing based on an operation state of the engine;
An ignition timing calculating means for correcting the basic ignition timing by an ignition timing correction amount for each cylinder when the engine shifts to an idle state; and an ignition means for igniting an air-fuel mixture of a corresponding cylinder based on an output of the ignition timing calculating means. Characteristic engine idle control device.