JP7560987B2 - Engine Control Unit - Google Patents

Description

The present invention relates to an engine control device that controls an engine having a direct fuel injection device.

For example, in gasoline engines installed in automobiles, it is important to precisely control the air-fuel ratio in order to stabilize combustion and control exhaust gas emissions.
For this reason, it is necessary to learn, for example, variations in components such as injectors and the concentration of purge gas (evaporative) introduced from the fuel evaporative emission control device to the engine intake system, and to reflect these in fuel injection control.

As a conventional technique related to learning correction for improving the control accuracy of the air-fuel ratio, for example, Patent Document 1 describes that in order to ensure the accuracy of air-fuel ratio control in the lean combustion mode, the conditions for switching from the stoichiometric combustion mode to the lean combustion mode include having performed injection amount learning in the stoichiometric combustion mode and having performed evaporative concentration learning in the stoichiometric combustion mode.
Patent Document 2 describes that in an engine having a main fuel injector that injects fuel into the combustion chamber and an auxiliary fuel injector that injects fuel into the intake passage, when the auxiliary fuel injector is switched between operating and non-operating, air-fuel ratio learning is prohibited for a predetermined period of time to prevent erroneous learning due to an air-fuel ratio error.
Furthermore, as a conventional technique for fuel injection control in a spark ignition direct injection engine, Patent Document 3 describes stratified combustion in which a homogeneous mixture that is relatively leaner than stoichiometric is formed throughout the combustion chamber by fuel injection during the intake stroke, and a mixture that is relatively richer than stoichiometric is formed around the spark plug by fuel injection during the compression stroke, and then burned.
Patent Document 4 describes a method of generating a tumble flow in a cylinder having an ignition plug at the top center, and injecting fuel toward the high-speed portion of the tumble flow when the high-speed portion moves toward the in-cylinder injector, thereby achieving good stratified charge combustion.

JP 2020-33933 A Japanese Patent Application Publication No. 11-351011 Japanese Patent Application Publication No. 11-324765 Patent Publication 2016-130495

During lean combustion, when an assist injection is performed to create a fuel-rich atmosphere around the spark plug in order to ensure ignition and combustion stability in addition to the main injection that is mainly for achieving the required air-fuel ratio, the required assist injection amount may not be accurately achieved because there are individual differences in the injection amount characteristics with respect to the injection signal applied to the injector. In such a case, the fuel required to ensure ignition cannot be injected, which may cause misfire.
In order to achieve stable lean combustion, it is necessary to learn the injection amount characteristics of the injector in assist injection, in which the fuel injection amount is very small, and to suppress the variation in the injection amount.
Typically, the injection amount characteristics of the injector during assist injection are learned during stoichiometric combustion at the initial start, for example, when a new vehicle is manufactured, or after the ECU memory is cleared by removing the battery.

On the other hand, in addition to the injection quantity characteristics of the injector in the assist injection described above, it is also necessary to learn the injection quantity characteristics of the injector in the main injection, which has a relatively large fuel injection quantity, and the concentration of the evaporative gas introduced from the canister of the fuel evaporative gas treatment device.
In any of these systems, the deviation of the actual air-fuel ratio detected by an air-fuel ratio sensor provided in the exhaust system from the required air-fuel ratio is fed back to the injection signal for the main injection.
In lean combustion, more precise control of the air-fuel ratio is generally required than in stoichiometric combustion, so that each of the above-mentioned learning processes needs to be performed with high accuracy.
To ensure accuracy, these learning processes cannot be performed simultaneously, but must be performed sequentially in chronological order. However, it takes time to learn the injection quantity characteristics of the injector for main injection and the concentration of evaporative gas. If the period until the injection quantity characteristics of the injector for assist injection are learned is long, the state in which lean combustion cannot be performed will be prolonged, and the vehicle's fuel efficiency will deteriorate.
In view of the above-mentioned problems, an object of the present invention is to provide an engine control device that can execute lean combustion early while ensuring the control accuracy of the air-fuel ratio.

The present invention solves the above-mentioned problems by the following solving means.
According to one aspect of the present invention, there is provided an engine control device for controlling an engine equipped with a fuel injection device having an injector that performs a first injection of fuel into a combustion chamber and a second injection of fuel that is smaller than the first injection after the first injection to form a mixture around an ignition plug that is richer in fuel than the mixture formed by the first injection, the engine control device including an injection amount learning unit that performs a first learning of learning an injection amount characteristic of the injector when performing the first injection and a second learning of learning an injection amount characteristic of the injector when performing the second injection, the fuel injection device forms a homogenous mixture in the combustion chamber by the first injection, and then forms a mixture near the spark plug that is richer in fuel than the surroundings by the second injection, and the average air-fuel ratio of the combustion chamber is reduced by a fuel-rich amount with respect to the stoichiometric air-fuel ratio. the injection amount learning unit determines whether or not there is a history of the second learning being performed a predetermined number of times or more after a memory is cleared, and changes a condition for determining the end of the first learning when the first learning is performed in an incomplete state indicating that there is no history of the second learning being performed a predetermined number of times or more to a condition for determining the end of the first learning when the first learning is performed in a completed state indicating that there is a history of the second learning being performed a predetermined number of times or more so that the end of the first learning is earlier, the second learning is performed after the end of the first learning performed when the second learning is in the incomplete state, and the fuel injection device prohibits the execution of the lean burn control when the second learning is in the incomplete state .
According to the present invention, when the second learning related to the injection amount characteristics when injecting a small amount of fuel is incomplete, the conditions for determining the end of the first learning are changed so that the end of the first learning is brought forward, thereby accelerating the time when both the first learning and the second learning are completed, thereby ensuring the accuracy of the fuel injection amount of the second injection and accelerating the time when lean combustion can be performed.

In the present invention, the engine is equipped with a fuel evaporative gas treatment device that introduces temporarily stored fuel evaporative gas into an intake system, and a fuel evaporative gas concentration learning unit that performs third learning to learn the concentration of the fuel evaporative gas introduced from the fuel evaporative gas treatment device, and the injection amount learning unit can be configured to perform the second learning in priority to the third learning when the second learning is incomplete .
According to this, when the second learning is incomplete, the second learning is executed in preference to the third learning, thereby making it possible to more reliably advance the time when lean combustion can be performed.
For example, the second learning can be performed in principle with priority over the third learning, and the third learning can be given priority over the second learning only when there is an urgency to perform the third learning due to, for example, an excessive amount of adsorption in the canister.

In the present invention, the injection amount learning unit can be configured to perform the first learning based on a correlation between an injection time of the injector and an air-fuel ratio in the first injection, and to perform the second learning based on a correlation between a drive voltage of the injector and an air-fuel ratio in the second injection.
This makes it possible to perform learning that appropriately reflects the respective injection amount characteristics in the region of injection amount used in the first injection and the region of injection amount used in the second injection, thereby improving the accuracy of air-fuel ratio control.

In the present invention, the fuel injection device has a function of performing the second injection under conditions in which the fuel pressure supplied to the injector is different, and the second learning can be configured to be performed for each of a plurality of different levels of fuel pressure.
This makes it possible to ensure the accuracy of air-fuel ratio control even when the pressure of fuel supplied to the injector changes.

As described above, the present invention provides an engine control device that can perform lean combustion early while ensuring the control accuracy of the air-fuel ratio.

1 is a diagram showing a schematic configuration of an engine having an embodiment of an engine control device to which the present invention is applied; FIG. 4 is a diagram illustrating an example of an injection amount characteristic of an injector according to the embodiment. 4 is a flowchart showing a learning process related to air-fuel ratio control in the engine control device according to the embodiment. 5 is a timing chart showing the implementation states of base air-fuel ratio learning, evaporative concentration learning, and pre-Q _min learning in the embodiment.

Hereinafter, an embodiment of an engine control device to which the present invention is applied will be described.
The engine control device of the embodiment generally controls a horizontally opposed four-cylinder gasoline direct injection turbocharged engine mounted as a power source for driving an automobile such as a passenger car, and its accessories.

FIG. 1 is a diagram showing a schematic configuration of an engine having an engine control device according to an embodiment of the present invention.
The engine 1 is configured to include a crankshaft 10, a cylinder block 20, a cylinder head 30, a turbocharger 40, an intake system 50, an exhaust system 60, a canister 70, an EGR device 80, an engine control unit (ECU) 100, and the like.

The crankshaft 10 is a rotating shaft that serves as the output shaft of the engine 1 .
One end of the crankshaft 10 is connected to a power transmission mechanism such as a transmission (not shown).
A piston is connected to the crankshaft 10 via a connecting rod (not shown).
A crank angle sensor 11 is provided at the end of the crankshaft 10 to detect the angular position of the crankshaft.
The output of the crank angle sensor 11 is transmitted to the ECU 100 .

The cylinder block 20 is configured as two halves that sandwich the crankshaft 10 from the left and right when the crankshaft 10 is mounted vertically on the vehicle body.
A crankcase portion is provided in the center of the cylinder block 20, which houses the crankshaft 10 and has main bearings that rotatably support the crankshaft 10.
Inside the left and right banks of the cylinder block 20 arranged on the left and right sides of the crankcase portion, for example, a pair of cylinders (in the case of a four-cylinder engine) are formed, each having a piston inserted therein that reciprocates therein.

The cylinder heads 30 are provided at the ends (left and right ends) of the cylinder block 20 opposite the crankshaft 10 .
The cylinder head 30 is equipped with a combustion chamber 31, an ignition plug 32, an intake port 33, an exhaust port 34, an intake valve 35, an exhaust valve 36, an intake camshaft 37, an exhaust camshaft 38, etc.
The combustion chamber 31 is formed by recessing a portion of the cylinder head 30 opposite to the piston crown surface, for example, in a pent roof shape.
The spark plug 32 is provided in the center of the combustion chamber 31 and generates a spark in response to an ignition signal from the ECU 100 to ignite the air-fuel mixture.

The intake port 33 is a flow passage that introduces combustion air (fresh air) into the combustion chamber 31 .
The exhaust port 34 is a flow passage for discharging burnt gas (exhaust gas) from the combustion chamber 31 .
The intake valve 35 and the exhaust valve 36 open and close the intake port 33 and the exhaust valve 34 at a predetermined valve timing.
For example, two intake valves 35 and two exhaust valves 36 are provided for each cylinder.
The intake valve 35 and the exhaust valve 36 are opened and closed by an intake camshaft 37 and an exhaust camshaft 38 which rotate synchronously at half the rotation speed of the crankshaft 10 .
The cam sprocket portions of the intake camshaft 37 and the exhaust camshaft 38 are provided with variable valve timing mechanisms (not shown) that advance or retard the phase of each camshaft to change the opening and closing timings of each valve.

The turbocharger 40 is a supercharger that utilizes the energy of the exhaust gas from the engine 1 to compress and supercharge combustion air (fresh air).
The turbocharger 40 includes a turbine 41, a compressor 42, an air bypass passage 43, an air bypass valve 44, a wastegate passage 45, a wastegate valve 46, and the like.
The turbine 41 is rotationally driven by exhaust gas from the engine 1 .
The compressor 42 is attached coaxially to the turbine 41 and is rotationally driven by the turbine 41 to compress air.

The air bypass passage 43 extracts a portion of the air from the downstream side of the compressor 42 and returns the air to the upstream side of the compressor 42 .
The air bypass valve 44 is provided in the air bypass flow path 43, and switches between two states in response to a command from the ECU 100: a closed state in which the air bypass flow path 43 is substantially blocked, and an open state in which air can pass through the air bypass flow path 43.
The air bypass valve 44 is an electric valve having a valve body that is driven to open and close by an electric actuator.
The air bypass valve 44 is opened to prevent surging of the turbocharger 40 and to protect the blades, for example when the throttle valve 56 is suddenly closed, and causes the air in the intake pipe downstream of the compressor 42 to flow back to the upstream side of the compressor 42, thereby reducing excess pressure.

The wastegate passage 45 extracts a portion of the exhaust gas from the upstream side of the turbine 41 and bypasses it to the downstream side of the turbine 41 for the purpose of controlling the boost pressure, raising the temperature of the catalyst, etc.
The wastegate passage 45 is integrally formed with the housing of the turbine 41 .
The wastegate valve 46 is provided in the wastegate passage 45 and has a valve body that opens and closes the passage, and controls the flow rate of exhaust gas passing through the wastegate passage 45.
The wastegate valve 46 is an electric wastegate valve having an electric actuator that drives a valve body to open and close in response to commands from the ECU 100 .
The wastegate valve 46 can be switched between a fully open state and a fully closed state, and can also be set to any intermediate position between these states.

The intake system 50 introduces air into the intake port 33 .
The intake system 50 includes an intake duct 51, a chamber 52, an air cleaner 53, an air flow meter 54, an intercooler 55, a throttle valve 56, an intake manifold 57, an intake pressure sensor 58, an injector 59, and the like.

The intake duct 51 is a flow path that introduces outside air into the intake port 33 .
The chamber 52 is a space provided in communication with the vicinity of the inlet of the intake duct 51 .
The air cleaner 53 is provided downstream of the portion of the intake duct 51 that communicates with the chamber 52, and filters the air to remove dust and the like.
The air flow meter 54 is provided near the outlet of the air cleaner 53 and measures the flow rate of air passing through the intake duct 51 .
The output of the air flow meter 54 is transmitted to the ECU 100 .
The compressor 42 of the turbocharger 40 is provided downstream of the air flow meter 54 .

The intercooler 55 is provided in the intake duct 51 downstream of the compressor 42, and is a heat exchanger that cools compressed and heated air by, for example, exchanging heat with airflow caused by running.
The throttle valve 56 is a butterfly valve that is provided in the intake duct 51 downstream of the intercooler 55 and adjusts the flow rate of air to control the output of the engine 1 .
The throttle valve 56 is driven to open and close by a throttle actuator (not shown) in response to the driver's operation of an accelerator pedal (not shown) or the like.
The throttle valve 56 is provided with a throttle sensor for detecting the opening degree thereof, and the output thereof is transmitted to the ECU 100 .
The intake manifold 57 is provided downstream of the throttle valve 56 and is a branch pipe that distributes air to the intake ports 33 of each cylinder.
The intake pressure sensor 58 detects the pressure of the air in the intake manifold 57 (intake pressure).
The output of the intake pressure sensor 58 is transmitted to the ECU 100 .
The injector 59 is provided at the end of the intake manifold 57 on the cylinder head 30 side, and injects fuel into the combustion chamber 31 in response to an injection signal issued by the ECU 100 to form an air-fuel mixture.

The exhaust system 60 discharges the exhaust gas discharged from the exhaust port 34 to the outside.
The exhaust system 60 includes an exhaust manifold 61, an exhaust pipe 62, a three-way catalyst 63, a storage reduction catalyst 64, a silencer 65, air-fuel ratio sensors 66 and 67, and the like.
The exhaust manifold 61 is a collecting pipe that collects the exhaust gases emitted from the exhaust ports 34 of the cylinders.
The turbine 41 of the turbocharger 40 is disposed downstream of the exhaust manifold 61 .
The exhaust pipe 62 is a pipe through which exhaust gas discharged from the turbine 41 is discharged to the outside.

The three-way catalyst 63 is provided in the middle part of the exhaust pipe 62 .
The three-way catalyst 63 purifies HC, NOx, CO, and the like contained in the exhaust gas.
The three-way catalyst 63 is provided adjacent to the outlet of the turbine 41 .
The three-way catalyst exerts a purification function within a predetermined active range in which the air-fuel ratio is close to the theoretical air-fuel ratio (stoichiometric).

The storage reduction catalyst 64 is provided in the middle of the exhaust pipe 62 and on the downstream side (outlet side) of the front catalyst 63 .
The rear catalyst 64 is a lean knock trap catalyst (LNT) that temporarily stores _NOX in the exhaust gas when the engine 1 is operated at a lean air-fuel ratio, and reduces _NOX using fuel as a reducing agent when the engine 1 is operated at a rich air-fuel ratio.
At the inlet and outlet of the storage reduction catalyst 64, a _NOx sensor (not shown) is provided for detecting the _NOx concentration in the exhaust gas.

The silencer 65 is provided near the outlet of the exhaust pipe 62 to reduce the acoustic energy of the exhaust gas.
The air-fuel ratio sensor 66 is provided between the outlet of the turbine 41 and the inlet of the three-way catalyst 63 .
The air-fuel ratio sensor 67 is provided between the outlet of the three-way catalyst 63 and the inlet of the storage reduction catalyst 64 .
The air-fuel ratio sensor 66 and the rear _O2 sensor 67 are both linear output sensors that detect the amount of oxygen in the exhaust gas by generating an output voltage corresponding to the oxygen concentration in the exhaust gas.
The outputs of the air-fuel ratio sensors 66 and 67 are both transmitted to the ECU 100 .

The canister (charcoal canister) 70 is a fuel evaporation gas treatment device into which fuel evaporation gas (evaporation gas) generated in a fuel tank (not shown) that stores gasoline used as fuel for the engine 1 is introduced and temporarily stored.
The canister 70 is configured by accommodating activated carbon capable of temporarily absorbing fuel evaporative gas in a canister case, which is a housing made of resin.

The canister 70 is configured with a purge line 71, which is mainly used when not supercharging, a purge control valve 72, and a purge line 73, which is mainly used when supercharging, a purge control valve 74, etc.

The purge line 71 is a flow passage whose two ends are connected to the canister 70 and the intake manifold 57, respectively, and which connects the interiors of these components.
The purge line 71 introduces purge gas, which is made up of fuel vapor released from the canister 70, into the intake manifold 57 when the pressure inside the intake manifold 57 is negative and the engine is not being supercharged.
The purge control valve (PCV) 72 is a duty-controlled solenoid valve provided in the middle of the purge line 71 .
The PCV 72 can be switched between an open state and a closed state and the opening degree in the open state can be set in response to a command from the ECU 100 .

The purge line 73 is a flow passage whose two ends are connected to the canister 70 and a region of the intake duct 51 adjacent to the inlet of the compressor 42, and which connects the interiors of these.
The purge line 73 introduces the purge gas into the intake duct 51 upstream of the compressor 42 during supercharging when the inside of the intake manifold 57 becomes positive pressure and it becomes difficult to introduce the purge gas through the purge line 71 .
The purge control valve (PCV) 74 is an electromagnetic valve provided midway along the purge line 73 .
The PCV 74 can be switched between an open state and a closed state in response to a command from the ECU 100 .

The EGR device 80 introduces (recirculates) exhaust gas extracted from the exhaust device 60 into the intake manifold 57 for the purposes of reducing pumping losses during partial load, reducing cooling losses by suppressing the combustion temperature, and suppressing the generation of _NOx , for example.
The EGR device 80 includes an EGR line 81, an EGR valve 82, an EGR cooler 83, and the like.

The EGR line 81 is a pipe that introduces exhaust gas from a part of the exhaust gas flow path to the intake manifold 57 .
In the example shown in FIG. 1 , the EGR line 81 extracts exhaust gas from the exhaust pipe 62 , but it may also be configured to extract the exhaust gas from the exhaust manifold 61 or the exhaust port 34 .
In response to commands from ECU 100, EGR valve 82 switches between an open state in which EGR gas (exhaust gas) can flow through EGR line 81 and a closed state in which EGR line 81 is blocked, and is capable of adjusting the opening degree (exhaust gas flow rate) in the open state.
The EGR cooler 83 is provided in the EGR line 81 and cools the exhaust gas by exchanging heat with, for example, the cooling water of the engine 1 or the wind generated when the engine is running.

An engine control unit (ECU) 100 comprehensively controls the engine 1 and its auxiliaries.
The ECU 100 is configured to include an information processing means such as a CPU, storage means such as a RAM and a ROM, an input/output interface, and a bus connecting these.
The ECU 100 is also provided with an accelerator pedal sensor 101 that detects the amount of depression of an accelerator pedal (not shown) by the driver.
The ECU 100 has a function of setting a driver requested torque based on the output of an accelerator pedal sensor 101 and the like.
The ECU 100 controls the throttle valve opening, the boost pressure, the fuel injection amount, the ignition timing, the valve timing, etc. so that the torque actually generated by the engine 1 approaches the set driver requested torque.

In addition, the ECU 100 has the function of switching between operation with stoichiometric combustion, in which the air-fuel ratio in the combustion chamber 31 is close to the theoretical air-fuel ratio, and operation with lean combustion, in which a fuel-lean atmosphere is created compared to stoichiometric combustion, depending on the load state of the engine.
In stoichiometric combustion, a homogeneous mixture is formed in the combustion chamber 31 by a single injection (main injection) during the intake stroke, for example.
In lean combustion, for example, a mixture that is homogeneous and leaner in fuel than stoichiometric combustion is formed in the combustion chamber 31 by injection during the intake stroke (main injection), and a mixture that is relatively richer in fuel around the spark plug 32 than other areas is formed by injection of a smaller amount than the main injection during the compression stroke (assist injection), thereby performing stratified combustion.

The ECU 100 has a function of performing base air-fuel ratio learning (first learning) to learn the correlation between the fuel injection amount and the injection pulse width (equivalent to the injection time) of the pulsed voltage (injection signal) supplied to the injector 59 in the range of the fuel injection amount used in the main injection.
In addition, the ECU 100 has a function of performing _Qmin learning (second learning) to learn the correlation of the fuel injection amount with respect to the voltage of the injection signal in the region of the fuel injection amount used in assist injection (the region near the minimum fuel injection _amount Qmin of the injector 59).
In actual fuel injection control, the fuel injection amount in the region near the minimum fuel injection amount _Qmin of the injector 59 (the region where the injector 59 does not open to the full open state) is expressed by a virtual injection pulse width required to inject the corresponding fuel injection amount, assuming that the injector 59 opens to the full open state.
Therefore, in _Qmin learning, a configuration can be adopted in which the correlation between a virtual injection pulse width (injection time) and the voltage of the injection signal is learned.
Among the Q _min learning, the one that is performed after the RAM of the ECU 100 is reset (memory cleared) and before the first lean combustion is permitted will be referred to as pre-Q _min learning hereinafter.

FIG. 2 is a diagram illustrating an example of an injection amount characteristic of the injector according to the embodiment.
In FIG. 2, the horizontal axis indicates the pulse width of the injection signal, and the vertical axis indicates the fuel injection amount.
As shown in FIG. 2, in a target range of base air-fuel ratio learning where the injection amount is relatively large, the fuel injection amount exhibits a characteristic of increasing approximately linearly in response to an increase in the pulse width of the injection signal.
This shows that in the region of main injection, the fuel injection amount can be controlled by the pulse width, that is, the injection time.
On the other hand, in the pre-Q _min learning region where the injection amount is relatively small, the correlation of the fuel injection amount with respect to the pulse width of the injection signal is different from that in the base air-fuel ratio learning region.
In such a region where the fuel injection amount is small, the ECU 100 controls the fuel injection amount by the voltage of the injection signal.
Between the base air-fuel ratio learning region and the pre-Q _min learning region is a bounce region in which it is difficult to precisely control the fuel injection amount.
The bounce region is not used during normal operation of the engine 1.

FIG. 3 is a flowchart showing a learning process related to air-fuel ratio control in the engine control device according to the embodiment.
Each step will be explained in order below.

<Step S01: Determining Pre-Q _min Learning Incompletion>
After resetting the RAM (clearing the memory), the ECU 100 determines whether or not there is a history of pre-Q _min learning having been performed a predetermined number of times or more.
If the pre-Q _min learning has been completed a predetermined number of times (lean combustion is permitted), the process proceeds to step S02. If the pre-Q _min learning has not been completed a predetermined number of times (lean combustion is not permitted), the process proceeds to step S05.

<Step S02: Lean combustion permission>
The ECU 100 sets the lean combustion to a permitted state.
This makes it possible to switch from stoichiometric combustion to lean combustion accompanied by assist injection depending on the operating state of the engine 1.
Then, proceed to step S03.

<Step S03: Execute base air-fuel ratio learning (normal)>
The ECU 100 executes the base air-fuel ratio learning when a predetermined base air-fuel ratio learning condition is satisfied.
In the base air-fuel ratio learning, for example, when the combustion is in a stoichiometric combustion state and in a steady state, an initial injection signal pulse width is set for a predetermined target air-fuel ratio, injection is performed, and the air-fuel ratio is detected by the air-fuel ratio sensor 66.
Thereafter, feedback for correcting the pulse width of the injection signal in accordance with the deviation between the detected actual air-fuel ratio and the target air-fuel ratio is repeated until a predetermined learning end time has elapsed.
During the execution of the base air-fuel ratio learning, the purge control valves 72, 74 are closed and the introduction of evaporative fuel is stopped.
After the learning end time has elapsed, the base air-fuel ratio learning is ended, and the process proceeds to step S04.

<Step S04: Execute Evaporative Particle Concentration Learning (Normal)>
The ECU 100 executes the evaporative gas concentration learning (third learning) when a predetermined evaporative gas concentration learning condition is satisfied.
In the evaporative concentration learning, for example, when the combustion is at stoichiometric combustion and in a steady state, the purge control valves 72, 74 are opened and closed in a predetermined pattern, and the fluctuation in the air-fuel ratio detected by the air-fuel ratio sensor 66 at that time is detected.
Thereafter, feedback for correcting the pulse width of the injection signal for the main injection in accordance with the deviation between the detected actual air-fuel ratio and the target air-fuel ratio is repeated until a predetermined learning end time has elapsed.
The amount of correction of the pulse width at this time reflects the concentration of the evaporative fuel supplied from the canister 70 .
After the learning end time has elapsed, the evaporative gas concentration learning ends, and the series of processing steps ends (returns).

<Step S05: Execute base air-fuel ratio learning (shortening)>
When a predetermined base air-fuel ratio learning condition is satisfied, the ECU 100 executes the base air-fuel ratio learning in the same manner as in step S03.
However, the learning end time at this time is shortened compared to normal times. The shortening amount of the learning end time can be set, for example, taking into consideration that the learning progresses to a minimum extent that the learning accuracy of the base air-fuel ratio learning does not pose a problem when lean combustion is performed.
After the shortened learning end time has elapsed, the base air-fuel ratio learning is ended, and the process proceeds to step S06.

<Step S06: Determining Urgency of Evaporative Particle Concentration Learning>
The ECU 100 determines whether there is a high need to purge the canister 70 (release the evaporative fuel) relative to the need to perform lean combustion, and therefore whether there is an urgency to execute the evaporative fuel concentration learning.
For example, when the amount of fuel evaporative gas stored in the canister 70 is equal to or greater than a predetermined upper limit and there is a risk of the evaporative gas being released outside the vehicle, it is determined that there is an urgency to learn the evaporative gas concentration.
If there is an urgency in learning the evaporative emission concentration, the process proceeds to step S10, and otherwise the process proceeds to step S07.

<Step S07: Execute pre-Qmin _learning >
The ECU 100 executes the pre-Q _min learning.
In the pre-Q _min learning, the fuel injection amount of the main injection is fixed while the engine 1 is operating at stoichiometric combustion, and a relatively small amount of fuel equivalent to the assist injection during lean combustion is injected following the main injection.
The voltage of the injection signal when injecting this small amount of fuel is changed in stages, and the change in the air-fuel ratio detected by the air-fuel ratio sensor 66 is monitored during this process to learn the correlation between the voltage of the injection signal and the amount of fuel injected.
When the pressure of the fuel supplied to the injector 59 (fuel pressure) is variable depending on the operating state of the engine 1, the pre-Q _min learning can be performed at each of a plurality of different fuel pressure levels.
After a predetermined learning end time has elapsed, the pre-Q _min learning is ended, and the process proceeds to step S08.

<Step S08: Execute Evaporative Concentration Learning (Shortened) >
When a predetermined evaporative gas concentration learning condition is satisfied, the ECU 100 executes the evaporative gas concentration learning in the same manner as in step S04.
However, the learning end time at this time is set to be shorter than that at the normal time. The shortening amount of the learning end time can be set by taking into consideration that the learning will proceed to a minimum extent that will not cause any problems in order to perform the canister purge.
This allows the evaporative concentration learning to be terminated early, and the purging of the canister 70 to be permitted.
After the learning end time has elapsed, the evaporative gas concentration learning ends, and the process proceeds to step S09.

<Step S09: Add Pre-Q _min Learning Count>
The ECU 100 increments the number of pre-Q _min learning operations.
Thereafter, the series of processes ends (returns).

<Step S10: Execute Evaporative Concentration Learning (Normal)>
When a predetermined evaporative gas concentration learning condition is satisfied, the ECU 100 executes the evaporative gas concentration learning in the same manner as in step S04.
After the learning end time has elapsed, the evaporative gas concentration learning ends, and the process proceeds to step S11.

<Step S11: Execute pre-Q _min learning>
The ECU 100 executes the pre-Q _min learning in the same manner as in step S07.
After a predetermined learning end time has elapsed, the pre-Q _min learning is ended, and the process proceeds to step S12.

<Step S12: Add Pre-Q _min Learning Count>
The ECU 100 increments the number of pre-Q _min learning operations.
Thereafter, the series of processes ends (returns).

FIG. 4 is a timing chart showing the implementation states of the base air-fuel ratio learning, the evaporative gas concentration learning, and the pre-Q _min learning in this embodiment.
FIG. 4(a) shows a normal state in which pre- _Qmin learning has been completed, and FIG. 4(b) shows a state in which pre-Qmin learning has not been completed.
As shown in Fig. 4A, once pre-Q _min learning is completed, there is no need to perform pre-Q _min learning again, and the base air-fuel ratio learning and the evaporative gas concentration learning can be performed until the normal learning end time has elapsed. This improves the learning accuracy of each learning, making it possible to perform more precise air-fuel ratio control.

In contrast, as shown in FIG. 4B, when the pre-Q _min learning is incomplete, the learning end time of the base air-fuel ratio learning is shortened and the pre-Q _min learning is executed with priority over the evaporative gas concentration learning. This makes it possible to quickly complete the pre-Q _min learning while maintaining the minimum necessary accuracy of the base air-fuel ratio learning to such an extent that, for example, misfire or combustion instability does not occur, and to advance the timing at which lean combustion is permitted, thereby improving the fuel efficiency of the vehicle.
Even if there is an urgency in learning the evaporative gas concentration and the evaporative gas concentration learning is performed prior to the pre-Q _min learning, the end time of the pre-Q _min learning can be shortened if the pre-Q min learning has not yet been completed, thereby bringing forward the end time of the pre-Q _min learning.

According to the present embodiment described above, the following effects can be obtained.
(1) In a state in which pre-Q _min learning, which is related to the injection amount characteristics when injecting fuel in the vicinity of the minimum fuel injection amount Q _min of the injector 59 used in assist injection, has not been completed, the conditions for determining the completion of the base air-fuel ratio learning related to the region used in main injection are changed so that the completion of the base air-fuel ratio learning is brought forward. This makes it possible to advance the timing at which both the base air-fuel ratio learning and the pre-Q _min learning are completed, ensure the accuracy of the fuel injection amount of the assist injection, and advance the timing at which lean combustion can be performed.
(2) When the pre-Q _min learning has not been completed, the pre-Q _min learning is executed with priority over the evaporative gas concentration learning, so that the time when lean combustion can be performed can be more reliably advanced.
(3) In the base air-fuel ratio learning, learning is performed based on the correlation between the injection time (pulse width) and the air-fuel ratio, and in the pre-Q _min learning, learning is performed based on the correlation between the injection signal voltage and the air-fuel ratio. As a result, it is possible to perform learning that appropriately reflects the injection amount characteristics in the injection amount range used in the main injection and in the injection amount range used in the assist injection, and the accuracy of the air-fuel ratio control can be improved.
(4) By performing pre-Q _min learning for each of a plurality of different levels of fuel pressure, the accuracy of the air-fuel ratio control of the assist injection can be ensured even when the fuel pressure supplied to the injector 59 changes.

(Modification)
The present invention is not limited to the above-described embodiment, and various modifications and variations are possible, which are also within the technical scope of the present invention.
(1) The configurations of the engine control device and the engine are not limited to the above-described embodiment and may be modified as appropriate.
For example, the engine cylinder layout, the number of cylinders, the presence or absence of a supercharger, the arrangement of catalysts and sensors, etc. can be changed as appropriate.
(2) In the embodiment, when the pre-Q _min learning is incomplete, the learning completion time of the base air-fuel ratio learning is shortened. However, the conditions for determining the completion of the base air-fuel ratio learning may be changed by a method other than the above so as to hasten the completion of the base air-fuel ratio learning.

REFERENCE SIGNS LIST 1 engine 10 crankshaft 11 crank angle sensor 20 cylinder block 30 cylinder head 31 combustion chamber 32 spark plug 33 intake port 34 exhaust port 35 intake valve 36 exhaust valve 37 intake camshaft 38 exhaust camshaft 40 turbocharger 41 turbine 42 compressor 43 air bypass flow path 44 air bypass valve 45 wastegate flow path 46 wastegate valve 50 intake system 51 intake duct 52 chamber 53 air cleaner 54 air flow meter 55 intercooler 56 throttle valve 57 intake manifold 58 intake pressure sensor 59 injector 60 exhaust system 61 exhaust manifold 62 exhaust pipe 63 three-way catalyst 64 storage reduction catalyst 65 Silencer 66, 67 Air-fuel ratio sensor 70 Canister 71 Purge line 72 Purge control valve 73 Purge line 74 Purge control valve 80 EGR device 81 EGR line 82 EGR valve 83 EGR cooler 100 Engine control unit (ECU)
101 Accelerator pedal sensor

Claims (4)

1. An engine control device for controlling an engine equipped with a fuel injection device having an injector that performs a first injection of fuel into a combustion chamber and a second injection of fuel that is smaller than the first injection after the first injection, thereby forming a fuel-rich mixture around an ignition plug than the mixture formed by the first injection,
an injection amount learning unit that performs first learning to learn an injection amount characteristic of the injector when performing the first injection and second learning to learn an injection amount characteristic of the injector when performing the second injection,
the fuel injection device forms a homogeneous mixture in the combustion chamber by the first injection, and then forms a mixture that is richer in fuel than the surroundings in the vicinity of the spark plug by the second injection, thereby executing lean combustion control in which an average air-fuel ratio of the combustion chamber is fuel-lean with respect to a stoichiometric air-fuel ratio,
The injection amount learning unit
determining whether or not there is a history of the second learning being performed a predetermined number of times or more after the memory is cleared;
a condition for determining whether the first learning is to be completed when the first learning is performed in an incomplete state, which indicates that there is no history of the second learning having been performed a predetermined number of times or more, is changed so that the first learning is completed earlier than a condition for determining whether the first learning is to be completed when the first learning is performed in a completed state, which indicates that there is a history of the second learning having been performed a predetermined number of times or more;
The second learning is performed after the first learning is completed while the second learning is in an incomplete state, and
the fuel injection device prohibits execution of the lean burn control when the second learning is incomplete;
An engine control device comprising: The engine includes a fuel evaporative gas treatment device that introduces temporarily stored fuel evaporative gas into an intake device;
a fuel evaporative gas concentration learning unit that performs third learning to learn the concentration of the fuel evaporative gas introduced from the fuel evaporative gas treatment device,
2. The engine control device according to claim 1 , wherein the injection amount learning section performs the second learning in priority to the third learning when the second learning is incomplete . 3. The engine control device according to claim 1, wherein the injection amount learning unit performs the first learning based on a correlation between an injection time of the injector and an air-fuel ratio in the first injection, and performs the second learning based on a correlation between a drive voltage of the injector and an air-fuel ratio in the second injection . the fuel injection device has a function of performing the second injection under a condition where a fuel pressure supplied to the injector is different,
4. The engine control device according to claim 1 , wherein the second learning is performed for each of a plurality of different fuel pressure levels.

Images