JP3233038B2 - Control device for in-cylinder injection spark ignition internal combustion engine - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for a direct injection spark ignition type internal combustion engine, and more particularly, to an idle control of a direct injection type spark ignition type internal combustion engine capable of injecting fuel in a compression stroke and an intake stroke. The present invention relates to a rotation speed control technology.

[0002]

2. Related Background Art In recent years, in a spark ignition type internal combustion engine mounted on a vehicle, fuel is directly supplied to a combustion chamber in place of a conventional intake pipe injection type in order to reduce harmful exhaust gas components and improve fuel efficiency. Various direct injection gasoline engines have been proposed. In a cylinder injection type gasoline engine, for example, by injecting fuel from a fuel injection valve into a cavity provided at the top of the piston, an air-fuel mixture having an air-fuel ratio close to the stoichiometric air-fuel ratio is generated around the ignition plug at the time of ignition Let me. This makes it possible to ignite even at a lean air-fuel ratio as a whole, thereby reducing CO and HC emissions and greatly improving fuel efficiency during idling and running under low load.

In such a gasoline engine,
The compression stroke injection mode (late injection mode) and the intake stroke injection mode (first injection mode) are switched according to the operating state of the engine, that is, the engine load.
As a result, during low-load operation, fuel can be injected during the compression stroke to form an air-fuel mixture having an air-fuel ratio close to the stoichiometric air-fuel ratio around the ignition plug or in the cavity. Good ignition can be realized even at a fuel ratio. On the other hand, during medium-high load operation, fuel can be injected during the intake stroke to form a mixture having a uniform air-fuel ratio in the combustion chamber, and thus, like the intake pipe injection type,
By burning a large amount of fuel, it is possible to secure the required output during acceleration or high-speed running.

[0004]

By the way, even in this in-cylinder injection type gasoline engine, the idling rotational speed is maintained stable during idling operation in the same manner as in a normal intake pipe injection type internal combustion engine. The idle rotation speed control is performed as much as possible. Normally, in this idle speed control, an air bypass valve (ABV) mainly functioning as an idle speed controller
, The valve position control is performed to correct the intake air amount. As a result, even when the idling rotational speed fluctuates, the idling rotational speed can always be stabilized.

However, even when the valve position of the air bypass valve is controlled as described above, a target intake air amount (target intake air amount) corresponding to the valve position control amount due to a temporal change of the air bypass valve or the like. May not match the actual intake air amount (actual intake air amount). Therefore,
The learning correction control of the valve position is performed for the purpose of preventing the deviation of the intake air amount due to the temporal change or the like.

However, in the cylinder injection type gasoline engine, the idling operation is performed in both the intake stroke injection mode and the compression stroke injection mode. Normally, the amount of intake air differs between the case of the intake stroke injection mode and the case of the compression stroke injection mode due to a difference in intake pressure during idling operation. For this reason, even if the learning in the intake stroke injection mode and the learning in the compression stroke injection mode are performed together, it is difficult to perform appropriate learning because there is a difference between the learning values. .

The present invention has been made in view of the above circumstances, and has as its object the in-cylinder injection type spark ignition which enables easy and proper learning control of the idle speed regardless of the fuel injection mode. An object of the present invention is to provide a control device for an internal combustion engine.

[0008]

In order to achieve the above object, according to the first aspect of the present invention, an intake stroke injection mode in which fuel is injected mainly in an intake stroke in accordance with an operation state, and a compression stroke mainly. In the control apparatus for a direct injection type spark ignition type internal combustion engine switchable to a compression stroke injection mode for performing fuel injection in the above, an idle state detection means for detecting an idle operation state of the internal combustion engine, and the idle state detection means is supplied to a combustion chamber. Intake air amount adjusting means for adjusting an intake air amount by adjusting an operating position; intake air amount detecting means for detecting an actual intake air amount adjusted by the intake air amount adjusting means and supplied to the combustion chamber; and the intake stroke injection. When the idle operation state is detected in one of the fuel injection mode of the compression stroke injection mode and the compression stroke injection mode, And operating position learning means for performing learning of the reference learned value of said operating position on the basis of the difference between the actual intake air amount information from the set target intake air amount and the intake amount detecting means Te, the one side fuel injection mode When the above criteria
While correcting the operating position using the learned value, but when the other side fuel injection mode, further comprising a correction means for correcting the operating position of the reference learned value conversion operation <br/> to Features.

Therefore, when the idle operation state is detected in one of the fuel injection mode of the intake stroke injection mode and the compression stroke injection mode, the target intake air amount and the intake air amount set according to the operation state. The learning of the reference learning value of the operating position of the intake air amount adjusting means is suitably performed based on the difference from the actual intake air amount information from the detecting means. When the fuel injection mode is on one side, the reference position of the operating position learning means
While the learning position is used as it is to correct the operating position,
In the other side fuel injection mode, the reference learning value is changed.
The operating position is corrected by the switching operation . That is, the base of the operating position by the operating position learning means in one side of the fuel injection mode.
The quasi-learning value is suitably reflected on the other fuel injection mode.

According to the second aspect of the present invention, the fuel injection mode is switched between an intake stroke injection mode in which fuel injection is performed mainly in an intake stroke and a compression stroke injection mode in which fuel injection is mainly performed in a compression stroke, according to an operating state. In a control device for a direct injection type spark ignition type internal combustion engine, an idle state detecting means for detecting an idle operation state of the internal combustion engine, and an intake air for adjusting an intake air amount supplied to a combustion chamber by adjusting an operation position. An amount adjustment unit, an intake amount detection unit configured to detect an actual intake air amount supplied to the combustion chamber adjusted by the intake amount adjustment unit, and one of the intake stroke injection mode and the compression stroke injection mode When the idle operation state is detected in the fuel injection mode, the target intake air amount set according to the operation state Learning performs the reference learned value of said operating position on the basis of the difference between the actual intake air amount information from the intake air quantity detecting means, when said idle operation state In the fuel injection mode of the other side is detected, once operation The operating position is learned based on the difference between the target intake air amount set according to the state and the actual intake air amount information from the intake amount detecting means, and the learning result of the other side is used as the reference learning of the one side. Operating position learning means for changing so as to correspond to the value and further learning a reference learning value of the operating position in the one side fuel injection mode, and using the reference learning value in the one side fuel injection mode. while correcting the operating position and,
In the other side fuel injection mode, the reference learning value
And a correcting means for correcting the operating position by converting the operation position.

Therefore, when the idling operation state is detected in one of the fuel injection mode of the intake stroke injection mode and the compression stroke injection mode, the target intake air amount and the intake air amount set according to the operation state. Learning of the reference learning value of the operating position of the intake air amount adjusting means is performed based on the difference from the actual intake air amount information from the detecting means. On the other hand, when the idling operation state is detected in the fuel injection mode on the other side, based on the difference between the target intake air amount once set according to the operation state and the actual intake air amount information from the intake amount detection means. Learning of the operating position is performed, and the learning result of the other side is changed to correspond to the reference learning value of the one side, whereby the learning of the reference learning value in the fuel injection mode of the one side is further performed. Therefore, the accuracy of the reference learning value is improved. In the fuel injection mode on one side, the operating position is corrected based on the reference learning value, and in the fuel injection mode on the other side, the operating position is corrected by converting the reference learning value . That is, the reference learning value of the operating position by the operating position learning means in one fuel injection mode is suitably reflected in the other fuel injection mode.

In the one-side fuel injection mode, the operating position is corrected based on the one-side learning result, and in the other-side fuel injection mode, the degree of correction based on the one-side learning result is changed. The working position is corrected as used. That is, the learning result of the operating position by the operating position learning means in the one side fuel injection mode is suitably reflected in the other side fuel injection mode.

[0013]

An embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a schematic configuration diagram showing one embodiment of a control device for an internal combustion engine according to the present invention mounted on a vehicle. Hereinafter, the configuration of the control device for the internal combustion engine will be described with reference to FIG.

The engine 1 is capable of performing fuel injection in the compression stroke (late injection mode) together with fuel injection in the intake stroke (first injection mode), and combustion at a lean air-fuel ratio, that is, a lean air-fuel ratio. In-cylinder in-line four-cylinder gasoline engine is applicable. In the in-cylinder injection type engine 1, an EGR device (exhaust gas recirculation device) that performs an intake device and an exhaust gas recirculation (EGR) including a combustion chamber.
Are designed exclusively for in-cylinder injection, and can easily be operated at a rich air-fuel ratio, a stoichiometric air-fuel ratio (stoichio) AFS, and a lean air-fuel ratio.

The cylinder head 2 of the engine 1 is also provided with an electromagnetic fuel injection valve 4 together with an ignition plug 3 for each cylinder, so that fuel is directly injected into the combustion chamber 5. On the top surface of the piston 7 slidably held by the cylinder 6, a hemispherical depression, that is, a cavity 8 is formed at a position where the fuel spray from the fuel injection valve 4 reaches in the latter half of the compression stroke. ing. The compression ratio of the engine 1 is set higher (for example, about 12) than that of the intake pipe injection type. The DOHC4 valve type is adopted as the valve operating mechanism.
In the upper part of the table, to drive the intake and exhaust valves 9 and 10, respectively,
An intake camshaft 11 and an exhaust camshaft 12 are rotatably supported.

The cylinder head 2 has both camshafts 1
An intake port 13 is formed in a substantially upright direction so as to pass through between the intake ports 1 and 12, and the intake flow passing through the intake port 13 is reversed in the combustion chamber 5 in the opposite direction to the normal tumble flow. It is possible to generate a tumble flow. On the other hand, the exhaust port 14 is formed in a substantially horizontal direction similarly to a normal engine, but a large-diameter exhaust gas recirculation port, that is, an EGR port 15 is branched obliquely downward.

In FIG. 1, reference numeral 16 denotes a water temperature sensor for detecting a cooling water temperature Tw. Reference numeral 17 denotes a vane type crank angle sensor that outputs a crank angle signal SGT at a predetermined crank position (for example, 5 ° BTDC and 75 ° BTDC) of each cylinder. The engine rotation speed Ne can be detected based on the Reference numeral 19 denotes an ignition coil that outputs a high voltage to the ignition plug 3. A camshaft that rotates at half the number of revolutions of the crankshaft is provided with a cylinder discrimination sensor (not shown) that outputs a cylinder discrimination signal SGC. It is identifiable whether or not it is.

A throttle body 23, a stepper motor type # 1 ABV (first air bypass valve) 24 functioning as an intake air amount correcting means 24, and an air flow sensor 32 are connected to the intake port 13 via an intake manifold 21 having a surge tank 20. And an intake pipe 25 provided with an air cleaner 22. The intake pipe 25 is provided with a large-diameter air bypass pipe 26 for bypassing the throttle body 23 and sucking air into the intake manifold 21. The large-diameter air bypass pipe 26 has a linear solenoid type large # 2ABV (second
An air bypass valve (27) is provided. The air bypass pipe 26 has a passage area similar to that of the intake pipe 25, and when the # 2 ABV 27 is fully opened, the engine 1
The required amount of intake in the low to medium speed range is possible.

The throttle body 23 has a butterfly type throttle valve 28 for opening and closing the flow path, and a throttle position sensor (a throttle valve opening sensor as a throttle valve opening sensor for detecting the opening of the throttle valve 28, that is, the throttle opening θth. Hereinafter, referred to as TPS) 29,
Detecting the fully closed state of the throttle valve 28 and detecting the engine 1
And an idle switch (idle state detection means) 30 for detecting the idling state of the vehicle. Actually, the TPS 29 outputs a throttle voltage Vth corresponding to the throttle opening θth, and the throttle opening θth is recognized based on the throttle voltage Vth.

The air flow sensor 32 detects the amount of intake air Qa, and for example, a Karman vortex flow sensor is used. This air flow sensor 32
Is capable of satisfactorily detecting a change in intake air amount based on the operation of # 1ABV24 or # 2ABV27. The intake air amount Qa may be obtained from a boost pressure sensor attached to the surge tank 20 and from the intake pipe pressure detected by the boost pressure sensor.

On the other hand, a three-way catalyst 42 and a muffler (not shown) are connected to the exhaust port 14 via an exhaust manifold 41 provided with an O ₂ sensor 40 capable of detecting an actual air-fuel ratio (actual A / F). The exhaust pipe 43 provided is connected. The above-mentioned EGR port 15 is connected to the upstream of the intake manifold 21 via a large-diameter EGR pipe 44, and a stepper motor type EGR valve 45 is provided in the pipeline.

The fuel tank 50 is installed at the rear of the vehicle (not shown). The fuel stored in the fuel tank 50 is sucked up by an electric low-pressure fuel pump 51 and fed to the engine 1 via a low-pressure feed pipe 52. The fuel pressure in the low-pressure feed pipe 52 is regulated to a relatively low pressure (low fuel pressure) by a first fuel pressure regulator 54 interposed in the return pipe 53. Engine 1
Supplied to the high-pressure feed pipe 5 by a high-pressure fuel pump 55 attached to the cylinder head 2.
The fuel is supplied to each fuel injection valve 4 through the delivery pipe 6 and the delivery pipe 57.

The high-pressure fuel pump 55 is, for example, of a swash plate axial piston type and is driven by the exhaust side camshaft 12 or the intake side camshaft 11 to generate a discharge pressure of 5 MPa to 7 MPa or more even when the engine 1 is idling. It is possible. And the delivery pipe 5
The fuel pressure in 7 is regulated to a relatively high pressure (high fuel pressure) by a second fuel pressure regulator 59 interposed in the return pipe 58.

In the figure, reference numeral 60 denotes the second fuel pressure regulator 5.
9 is an electromagnetic type fuel pressure switching valve attached to 9. The fuel pressure switching valve 60 is capable of relieving fuel in the ON state, thereby reducing the fuel pressure in the delivery pipe 57 to a low fuel pressure. Reference numeral 61 denotes a high-pressure fuel pump 55
A part of fuel used for lubrication and cooling of
This is a return pipe that returns to 0.

In the cabin of the vehicle, an input / output device, a storage device (RO) for storing a control program, a control map, and the like are provided.
M, RAM, BURAM, etc.), central processing unit (CP
U), an ECU (Electronic Control Unit) 70 including a timer counter and the like is installed, and the ECU 70 performs comprehensive control of the engine 1. ECU7
The various sensors described above are connected to the input side of 0, and detection information from these various sensors is input. E
The CU 70 determines the fuel injection mode, the fuel injection amount, the ignition timing, the EGR gas introduction amount, and the like based on the detected information, and determines the fuel injection valve 4, the ignition coil 19, the E
Drive control of the GR valve 45 and the like is performed. Although not described, the input side of the ECU 70 is connected to a large number of switches and sensors (not shown) in addition to the various sensors described above. Etc. are connected.

Next, the engine 1 configured as described above
The operation of the control device, that is, the outline of the combustion control will be described. When the engine 1 is in the cold state, when the driver turns on the ignition key, the ECU 70 turns on the low-pressure fuel pump 51 and the fuel pressure switching valve 60 to supply the fuel at a low fuel pressure to the fuel injection valve 4.

When the driver operates the ignition key, the engine 1 is cranked by a cell motor (not shown), and at the same time, the ECU 70 starts combustion control. At this time, the ECU 70 selects the first-stage injection mode (that is, the intake stroke injection mode) and injects fuel so as to have a relatively rich air-fuel ratio. This is based on the fact that since the fuel vaporization rate is low when the engine is cold, misfires and emission of unburned fuel (HC) cannot be avoided if the fuel is injected in the late injection mode (that is, the compression stroke injection mode). The ECU 70 closes the # 2ABV 27 at the time of such a start. Therefore, in this case, the intake air to the combustion chamber 5 is generated by the gap of the throttle valve 28 or the # 1 ABV
24. Note that # 1ABV24 and # 2A
The BV 27 is centrally managed by the ECU 70,
The respective valve opening amounts are determined according to the required amount of intake air (bypass air) bypassing the throttle valve 28.

When the start of the engine 1 is completed as described above and the engine 1 starts the idling operation, the high-pressure fuel pump 55 starts the rated discharge operation and the EC is started.
U 70 turns off the fuel pressure switching valve 60 and supplies high-pressure fuel to the fuel injection valve 4. At this time, the required fuel injection amount is obtained from the discharge pressure of the high-pressure fuel pump 55 and the valve opening time of the fuel injection valve 4, that is, the fuel injection time.

Until the cooling water temperature Tw rises to a predetermined value, the ECU 70 selects the first injection mode and injects fuel so as to obtain a rich air-fuel ratio as in the case of starting, and continues # 2ABV27. Closed state. When the engine 1 is warmed up, the ECU 70 sets the TPS 29
2 based on the target average effective pressure Pe obtained from the throttle opening information θth based on the throttle voltage Vth based on the throttle voltage Vth or the volumetric efficiency Ev calculated based on the intake air amount Qa from the air flow sensor 32 and the engine rotation speed Ne. The fuel injection mode area is searched from the fuel injection control map shown in FIG. Thereby, the fuel injection mode is set. Then, the fuel injection amount is determined according to the fuel injection mode, and the drive of the fuel injection valve 4 is controlled. At the same time, # 2
Open / close control of the ABV 27 and the EGR valve 45 is also performed.

For example, when the engine 1 is in a low-load / low-speed range such as during idling or low-speed running, the fuel injection mode is set to the late injection lean mode based on FIG. 2, and fuel injection is performed during the compression stroke. As well as
Based on the target average effective pressure Pe, the fuel injection amount is determined so as to achieve a lean target air-fuel ratio, that is, a target A / F (for example, A / F = about 30 to 40). At the same time, the ignition timing Sa,
The EGR amount Legr is set, and thereby good combustion control is performed.

The combustion in the late injection lean mode will be described in more detail. In the in-cylinder injection type engine 1, as described above, the cavity 8 is formed on the upper surface of the piston 7. From this, the intake port 13
Of the fuel and the intake air, that is, the fuel spray is injected into the ignition plug 3
Good concentration in the vicinity. As a result, a mixture near the stoichiometric air-fuel ratio AFS is always formed in a layer around the spark plug 3 at the time of ignition. Therefore, in the latter injection mode, good ignitability is ensured even at a lean air-fuel ratio as a whole.

Further, for example, when the engine 1 is in the middle load range as in the case of running at a constant speed, the fuel injection mode is set to the first injection lean mode or the stoichiometric feedback mode based on FIG. In these early injection lean mode and stoichiometric feedback mode, fuel injection is performed in the intake stroke. In the first-stage injection lean mode, the target A / F is a lean target A / F (for example, A / F) based on the volumetric efficiency Ev instead of the target average effective pressure Pe.
The fuel injection amount is determined so that F = about 20 to 23). At the same time, based on the volumetric efficiency Ev, the ignition timing Sa, EGR
The quantity Legr is set, which results in good combustion control.

On the other hand, in the stoichiometric feedback mode, the ignition timing Sa and the EGR amount Leg are determined based on the volumetric efficiency Ev.
r is set. In this case, the air-fuel ratio feedback control is performed according to the output voltage of the O ₂ sensor 40, and the target A / F is set to the stoichiometric air-fuel ratio AFS. Controlled. Further, for example, when the engine 1 is in a high load region such as during rapid acceleration or high-speed running, the fuel injection mode is set to the open loop mode based on FIG. 2, and in this case, the former injection mode is selected. The fuel injection is performed in the intake stroke, and the target A / F is set to a relatively rich air-fuel ratio based on the volumetric efficiency Ev as described above. Further, the ignition timing S
a, The EGR amount Legr is set, and good combustion control is performed.

When coasting during middle-to-high speed running, etc.,
The fuel injection mode is a fuel cut mode as shown in FIG. 2, in which case the fuel injection is stopped. This fuel cut is immediately stopped when the engine rotation speed Ne falls below the return rotation speed or when the driver depresses the accelerator pedal. By the way, when the engine 1 is in the idling operation state, in order to stabilize the idling operation, the idling rotational speed is adjusted in the same manner as in the case of the intake pipe injection type engine in accordance with the increase or decrease of the load of auxiliary equipment such as an air conditioner. You need to take control. In the control of the idling rotational speed, the opening degree control of the # 1 ABV 24, that is, the valve position P
The intake air amount is corrected by controlling obj.

However, as described above, # 1ABV2
Even if the valve position Pobj of No. 4 is controlled, a target intake air amount (hereinafter, referred to as a target intake air amount) according to the valve position control amount of the # 1 ABV 24 and an actual intake air amount due to a temporal change of the # 1 ABV 24 and the like. (Hereinafter referred to as the actual intake air amount) in some cases. Therefore, learning correction control of the valve position Pobj is performed so that the deviation of the intake air amount due to the above-mentioned temporal change is prevented, and the target intake air amount matches the actual intake air amount.

Incidentally, in the in-cylinder injection type engine 1, as described above, the idling operation may be performed in the first injection mode or in the second injection lean mode. The amount of intake air differs between the first injection mode and the second injection mode due to a difference in intake pressure during idling operation. in this way,
If the intake air amount during the idling operation is different, it is difficult to perform the learning control of the valve position Pobj together in the first injection mode and the second injection mode. For example, the air-fuel ratio is generally larger in the latter-stage injection mode than in the former-stage injection mode, and therefore, the required intake air amount during idling operation is larger. If the obtained learning value is applied to the previous injection mode as it is, the correction amount of the valve position Pobj is too large, and the idle rotation speed becomes too large inadvertently in the previous injection mode. Conversely, if the learning value obtained in the first injection mode is to be applied to the second injection mode as it is, in this case, the correction amount is too small,
In the latter injection mode, the learning effect is hardly exhibited.

Therefore, in the control apparatus for a direct injection type spark ignition type internal combustion engine of the present invention, the learning value of the valve position Pobj in one of the fuel injection modes of the first injection mode and the second injection mode is calculated. Only the reference learning value is learned as a reference, and the valve position Pobj in the other fuel injection mode is corrected by converting the reference learning value. Hereinafter, correction contents of the valve position Pobj will be described. Since the learning of the valve position Pobj is performed over a long period of time (long time), it is hereinafter referred to as long time learning.

Referring to FIG. 3, there is shown a flowchart of an idle speed control routine including a control for correcting the valve position Pobj according to the present invention, which is executed by the ECU 70. Hereinafter, based on FIG. And the long-time learning of the valve position Pobj of the # 1 ABV 24 will be described.

In step S10 in FIG. 3, it is determined whether or not the idle condition is satisfied and the engine 1 is in an idle operation state. Here, it is determined whether the idle switch 30 is on and the throttle valve 28 is fully closed. The determination result of step S10 is false (No)
In this case, the routine is exited without doing anything. on the other hand,
If the determination result of step S10 is true (Yes),
It can be determined that the engine 1 is in the idling operation state, and then the process proceeds to step S12.

In step S12, it is determined whether or not the engine speed Ne is currently being fed back (Ne-F / B). The state in which Ne-F / B is being performed means that there is no load change and the idle operation state is continued. If the result of the determination in step S12 is true and it is determined that Ne-F / B is currently in progress, the process proceeds to step S14.

In step S14, it is determined whether the conditions for performing the long time learning are satisfied. The condition for performing the long time learning is that a predetermined time (for example, 5 seconds) has elapsed since the result of the determination in step S12 is true. That is, in step S14, after the result of the determination in step S12 is true, it is determined whether or not the timer counter of the ECU 70 has counted a predetermined time (for example, 5 seconds).

If the decision result in the step S14 is false, the routine is repeatedly executed, and waits for a predetermined time (for example, 5 seconds) to elapse. On the other hand, step S1
If it is determined that the determination result of step 4 is true and that a predetermined time (for example, 5 seconds) has elapsed since the determination result of step S12 is true, the process proceeds to step S16. In step S16, it is determined whether or not the value of the timer TM which is set to the value N in the next step S20 is 0.

Immediately after the result of the determination in step S14 is true, the value of the timer TM is set to the initial value 0, as will be described later. The process proceeds to step S18. Step S18
Then, the timer TM is set to the value N. This value N is a value corresponding to a predetermined time ta (for example, 1 sec). For more information,
The product of the execution cycle Tc of the routine and the value N is a predetermined time ta.
(For example, 1 sec), the value N is set (Tc · N = ta).

The following step S20 is a step for performing long time learning (operating position learning means). Less than,
The learning procedure of the long time learning will be described. In step S20, it is determined whether or not the fuel injection mode is the late injection mode, that is, the late injection lean mode. If the determination result is true and the fuel injection mode is the late injection lean mode, the process proceeds to step S22.

In step S22, a long time learning value PL as the reference learning value is calculated from the following equation (1). PL = PL + KA · ΔP (1) Here, ΔP means a correction value (referred to as a real-time learning value) corresponding to the difference between the target intake air amount and the actual intake air amount. This is a value set in accordance with the difference between the target intake air amount calculated by the valve position control amount of 1 ABV 24 and the actual intake air amount detected by the air flow sensor 32. Further, KA is a long time learning value update gain in the late injection lean mode, and its value is, for example, 3/256.

That is, in this case, the long time learning value P
L is the long-time learning value update gain KA (for example, the previous long-time learning value PL and real-time learning value ΔP).
3/256). On the other hand, when the result of the determination in step S20 is false and the fuel injection mode is determined to be the former injection mode, the process proceeds to step S24.

In step S24, a long-time learning value PLK in the first-half injection mode is calculated from the following equation (2). PLK = KLK · (PL + KK · ΔP) (2) By the way, referring to this equation (2), the right side (PL + KK ·
The term ΔP) is substantially the same as the equation for calculating the long time learning value PL. That is, this (PL + KK · Δ
The term P) is obtained by adding the previous long-time learning value PL to the real-time learning value ΔP and the long-time learning value update gain KK.
(For example, 3/256) to calculate the long time learning value PL. That is,
In the first-stage injection mode, similarly to the case of the second-stage injection lean mode, first, the long-time learning value PL is obtained, and the long-time learning value PL is multiplied by a conversion coefficient (weight) KLK. The long time learning value PLK in the first injection mode is determined.

The long-time learning value update gain KK
Here, as in the case of the long time learning value update gain KA, for example, is set to 3/256, but it is not always necessary to match this. Also, the conversion coefficient K
LK is a value that is properly set in advance based on the difference in the intake air amount during the idle operation between the late injection lean mode and the first injection mode.

The long time learning value PLk in the first injection mode is calculated in this manner. In the next step S26, the long time learning value PL is newly obtained from the following equation (3). PL = PLK / KLK (3) That is, in the above equation (3), (PL + KK · ΔP) is properly set and stored as the long time learning value PL.
As a result, the long time learning value PL is favorably updated not only in the late injection lean mode but also in the first injection mode.

Referring to FIG. 4, the learning procedure of the long time learning described above is schematically shown. As described above, in the long time learning according to the present invention, the fuel injection mode is set to the former injection mode. In this case, the long-time learning value PLK is obtained, but a new long-time learning value PL is obtained at the same time (solid arrow) in the same manner as in the case of the late injection lean mode (broken arrow). That is, in this long-time learning, only the long-time learning value PL is learned as a single reference learning value for the long-time learning. The learning mode is also updated in the injection mode. As a result, the accuracy of the long time learning value PL is improved.

The routine of FIG. 3 is repeatedly executed.
If the determination result of step S14 is again true,
Next, step S16 is executed. However, this time, the value of the timer TM is not 0, and therefore, the determination result of step S16 is false, and in this case, the process proceeds to step S28. In step S28, the timer TM counts down (TM = TM-1). Then, the countdown of the timer TM proceeds, and the timer TM again reaches the value 0.
When the determination result of step S16 is true, the long time learning after step S20 is repeatedly executed. That is, the long time learning is performed by the timer T
Every time M counts down the value N, that is, the predetermined time ta
(For example, 1 second).

As described above, the long time learning value P
When L is learned, the valve position Pobj of # 1 ABV 24 is appropriately corrected based on the long time learning value PL. Hereinafter, the valve position Pobj of # 1ABV24
Will be described. If the result of the determination in step S12 is false, that is, Ne-F / B is not currently in effect, and the valve position Pobj
Is determined to be necessary, it is determined in the next step S30 whether or not the fuel injection mode is the late injection lean mode.

If the result of the determination in step S30 is true and the fuel injection mode is determined to be the late injection lean mode, the process proceeds to step S32, where the valve position Pobj is corrected based on the following equation (4) ( Correction means). Pobj = PA + PL (4) Here, PA indicates the reference valve position during the idling operation in the late injection lean mode.

That is, in the late injection lean mode, the long time learning value PL is used as it is.
On the other hand, if the result of the determination in step S30 is false and it is determined that the fuel injection mode is the first-stage injection mode, the process proceeds to step S34,
The valve position Pobj is corrected based on the following equation (5) (correction means).

Pobj = PB + KLK · PL (5) Here, PB indicates the reference valve position during idling operation in the first injection mode. The right side KL of this equation (5)
As is apparent from the above equation (3), the term of K · PL is nothing but the long-time learning value PLK in the previous injection mode. However, as shown in the same equation, in the first-stage injection mode, the long-time learning value PL, which is the reference correction value, is multiplied by the conversion coefficient KLK. The position Pobj is appropriately corrected.

In this manner, the valve position Pobj is appropriately corrected for each fuel injection mode, and the time period of # 1ABV24 is not affected regardless of whether the fuel injection mode is the first injection mode or the second injection lean mode. The target intake air amount always matches the actual intake air amount irrespective of the change. Therefore, the idle speed control, that is, the valve position control of the # 1 ABV 24 is made favorable,
It is possible to always maintain a stable idle rotation speed regardless of the fuel injection mode.

In step S36, the timer T is used as preparation for the next execution of the long time learning.
Reset M to initial value 0. In addition, idle operation,
Normally, since the engine 1 is frequently used after the warm-up is completed, it is often performed in the late injection lean mode. Therefore, in the above embodiment, the long time learning value PL in the late injection lean mode is mainly used as the reference learning value, and this long time learning value PL is reflected in the correction of the valve position Pobj in the previous injection mode. However, a good result is obtained even if learning is performed using the long time learning value PLK in the first injection mode as a reference learning value, and this long time learning value PLK is reflected in the correction of the valve position Pobj in the second injection lean mode. Is obtained.

As described in detail above, in the control device for the direct injection type spark ignition type internal combustion engine of the present invention, the fuel injection mode of one of the late injection mode and the early injection mode (for example, the late injection lean mode). The learning value (for example, the long time learning value PL) in the first mode is unified as the reference learning value to perform the long time learning of the valve position Pobj, and the one fuel injection mode (for example, the late injection lean mode) When correcting the valve position Pobj of the # 1 ABV 24 in the above, the reference learning value (for example,
The long-time learning value PL) is used as it is, and in the other fuel injection mode (for example, the previous injection mode), # 1
When correcting the valve position Pobj of the ABV 24, the reference learning value (for example, the long time learning value PL) is converted into an appropriate value using the conversion coefficient KLK and used (step S34 in FIG. 3). .

Therefore, by unifying the reference learning value of the long time learning, the learning value to be actually learned can be stabilized without variation, and the fuel injection can be performed by a simple operation of conversion. # 1 regardless of mode
The target intake air amount and the actual intake air amount can be suitably matched by always setting the valve position Pobj of the ABV 24 to an appropriate value. As a result, the idle speed control becomes good, and a stable idle speed can be maintained.

Further, in the control device for a direct injection type spark ignition type internal combustion engine of the present invention, the learning value (for example, the late injection lean mode) in one of the latter injection mode and the earlier injection mode is used. For example, the long-time learning value PL is used as a reference learning value, and the reference learning value (for example, the long-time learning value PL) is learned and updated also in the other fuel injection mode (for example, the previous injection mode). (Step S26 in FIG. 3). Therefore, the chance of learning the reference learning value (for example, the long time learning value PL) increases, and the accuracy of the reference learning value (for example, the long time learning value PL) can be improved.
As a result, the idle rotation speed control can be performed with the valve position Pobj of the # 1 ABV 24 being more appropriate, and a more stable idle rotation speed can be maintained.

[0061]

As described in detail above, claim 1 is as follows.
According to the control device of the in-cylinder injection spark ignition type internal combustion engine, when the idle operation state is detected in one of the fuel injection modes of the intake stroke injection mode and the compression stroke injection mode, the operation state is changed to the operation state. Based on the difference between the target intake air amount set accordingly and the actual intake air amount information from the intake amount detection means, the reference learning value of the operating position of the intake air amount adjusting means is suitably learned. Can be.

In the fuel injection mode on one side, the operating position is corrected by using the reference learning value of the operating position learning means as it is, while in the fuel injection mode on the other side, the reference learning value is corrected. The operating position can be properly corrected by the conversion operation . In other words, the operating position is determined by the operating position learning means in the fuel injection mode on one side.
Based on the reference learning value , this can be suitably reflected in the fuel injection mode on the other side.

As a result, the learning result is set as one reference learning value.
, The learning value for actual learning varies.
It can be stable without any matter, regardless of the fuel injection mode,
The learning correction control can be facilitated and made appropriate, and the idling operation state can always be kept stable. Further, according to the control device of the direct injection type spark ignition type internal combustion engine of the second aspect, the idle operation state is detected in one of the fuel injection modes of the intake stroke injection mode and the compression stroke injection mode. In some cases, based on the difference between the target intake air amount set according to the operation state and the actual intake air amount information from the intake air amount detection means, the criterion of the operating position of the intake air amount adjusting means is determined.
Learning of the learning value can be suitably performed.

On the other hand, even when the idling operation state is detected in the other fuel injection mode, the target intake air amount once set according to the operation state and the actual intake air amount information from the intake amount detecting means are determined. performs the learning of the operating position on the basis of the difference between, by properly changed to correspond to the learning results of the other side to a reference learned value of one side, the reference operating position in one side fuel injection mode The learning value can be further learned. Thereby, the accuracy of the reference learning value can be improved.

[0065] Then, when the one side fuel injection mode, while correcting the operating position so as to accept the reference learned value, when the other side fuel injection mode, based on
The operation position can be appropriately corrected by converting the quasi-learning value . That is, the reference learning value of the operating position by the operating position learning means in one fuel injection mode can be suitably reflected in the other fuel injection mode.

As a result, the learning result is set as one reference learning value.
, The learning value for actual learning varies.
It can be stable without any matter, regardless of the fuel injection mode,
The learning correction control can be facilitated and made appropriate, and the idling operation state can always be kept stable.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram showing a direct injection type spark ignition type internal combustion engine and a control device thereof.

FIG. 2 is a diagram showing a determination map of a fuel injection mode.

FIG. 3 is a flowchart showing a control routine of idle speed control according to the present invention.

FIG. 4 is a diagram schematically showing a learning procedure of long time learning in FIG. 3;

[Explanation of symbols]

Reference Signs List 1 engine 4 fuel injection valve 24 # 1ABV (first air bypass valve, intake air amount adjusting means) 29 TPS 30 idle switch (idle state detecting means) 32 air flow sensor (intake air amount detecting means) 70 electronic control unit (ECU)

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification code FI F02D 45/00 340 F02D 45/00 340H (72) Inventor Toshiro Nomura 5-33-8 Shiba 5-chome, Minato-ku, Tokyo Mitsubishi Motors Corporation (56) References JP-A-4-321747 (JP, A) JP-A-6-101547 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) F02D 41/00 -41/40

Claims (2)

(57) [Claims] An in-cylinder injection type spark which can switch a fuel injection mode between an intake stroke injection mode in which fuel is injected mainly in an intake stroke and a compression stroke injection mode in which fuel is injected mainly in a compression stroke, according to an operation state. A control device for an ignition type internal combustion engine, wherein: an idle state detection means for detecting an idle operation state of the internal combustion engine; an intake air amount adjustment means for adjusting an intake air amount supplied to a combustion chamber by adjusting an operation position; An intake air amount detecting means for detecting an actual intake air amount which is adjusted by an amount adjusting means and supplied to the combustion chamber; and a fuel injection mode on one of the intake stroke injection mode and the compression stroke injection mode. When the idling operation state is detected, the target intake air amount set according to the operation state and the actual intake air amount from the intake air amount detection means. And operating position learning means for performing learning of the reference learned value of said operating position on the basis of the difference between the air amount information, when the fuel injection mode of the one side, the reference learned value
And correcting means for correcting the operating position by converting the reference learning value when the other side is in the fuel injection mode. A control device for an injection-type spark ignition internal combustion engine. 2. An in-cylinder injection type spark capable of switching a fuel injection mode between an intake stroke injection mode in which fuel is injected mainly in an intake stroke and a compression stroke injection mode in which fuel is injected mainly in a compression stroke according to an operation state. A control device for an ignition type internal combustion engine, wherein: an idle state detection means for detecting an idle operation state of the internal combustion engine; an intake air amount adjustment means for adjusting an intake air amount supplied to a combustion chamber by adjusting an operation position; An intake air amount detecting means for detecting an actual intake air amount which is adjusted by an amount adjusting means and supplied to the combustion chamber; and a fuel injection mode on one of the intake stroke injection mode and the compression stroke injection mode. When the idling operation state is detected, the target intake air amount set according to the operation state and the intake air amount detection means Performs <br/> learning of the reference learned value of said operating position on the basis of the difference between the intake air amount information, when the idle operating state In the fuel injection mode of the other side is detected, once depending on operating conditions The operating position is learned based on the difference between the target intake air amount set in advance and the actual intake air amount information from the intake amount detecting means, and the learning result on the other side is used as the reference learning on the one side.
Operating position learning means for changing to correspond to the value and further learning a reference learning value of the operating position in the one side fuel injection mode; and in the one side fuel injection mode, the reference learning value
A correcting means for correcting the operating position by converting the reference learning value in the other-side fuel injection mode while correcting the operating position using Control device for internal injection type spark ignition type internal combustion engine.