JPH0450448Y2 - - Google Patents

Description

[Detailed description of the invention] <Industrial field of application> The present invention relates to an air-fuel ratio learning control device for an automobile internal combustion engine that has an electronically controlled fuel injection device with an air-fuel ratio feedback control function. The present invention relates to an air-fuel ratio learning control device that can respond well to changes.

<Prior art> Conventionally, in an internal combustion engine having an electronically controlled fuel injection device having an air-fuel ratio feedback control function, Japanese Patent Laid-Open No. 60-90944 and Japanese Patent Laid-Open No. 61-190142 have been disclosed.
An air-fuel ratio learning control device as shown in the above publication is employed.

This is based on the signal from the _O2 sensor installed in the engine exhaust system, which calculates the basic fuel injection amount from engine operating state parameters related to the amount of air taken into the engine (for example, engine intake air flow rate and engine speed). The standard value of the feedback correction coefficient during air-fuel ratio feedback control, in which the fuel injection amount is calculated by correcting it with a feedback correction coefficient set by proportional/integral control based on the air-fuel ratio, and the air-fuel ratio is feedback-controlled to the target air-fuel ratio. A learning correction coefficient is determined by learning the deviation from the area for each predetermined engine operating state area, and when calculating the fuel injection amount, the basic fuel injection amount is corrected by the area-specific learning correction coefficient and feedback is calculated. The base air-fuel ratio obtained from the fuel injection amount calculated without correction by the correction coefficient is made to match the target air-fuel ratio, and during air-fuel ratio feedback control, this is further corrected by the feedback correction coefficient to calculate the fuel injection amount. It is something to do.

According to this, it is possible to eliminate the follow-up delay of the feedback control during the transient operation during the air-fuel ratio feedback control, and it is possible to accurately obtain the desired air-fuel ratio when the air-fuel ratio feedback control is stopped.

In addition, a system that determines the basic fuel injection amount Tp from the throttle valve opening α and the engine speed N (for example, calculates the intake air flow rate Q from α and N with reference to a map, Tp = K・Q/N (K is a constant)
Alternatively, an air flow meter is used to detect the intake air flow rate Q, and based on this and the engine speed N, the basic fuel injection amount Tp=K・
In systems that calculate Q/N and use a flap type (volume flow rate detection type) air flow meter, changes in air density are not reflected in the calculation of the basic fuel injection amount. For example, as long as learning progresses satisfactorily, changes in air density due to altitude or intake temperature can be accommodated.

<Problem that the invention aims to solve> However, if we consider the case of suddenly climbing to a high altitude (mountain), we find that when climbing a mountain, there is a transient driving pattern, so the method of learning by area of engine operation is difficult to learn. The area for learning is not determined, and even if learning is possible, the area is limited, and in the majority of areas, learning hardly progresses. This results in
When you start driving normally on flat ground near the top of a mountain,
Due to the control delay in the air-fuel ratio feedback control, the base air-fuel ratio deviates significantly from the target air-fuel ratio when the air-fuel ratio feedback control is stopped, resulting in poor drivability.

This requires correcting changes in air density by learning from the deviation from the reference value of the feedback correction coefficient during air-fuel ratio feedback control, but some of the learned deviations include variations in parts such as fuel injectors and throttle bodies. etc., it also includes deviations in the base air-fuel ratio that depend on the engine operating state, so it is impossible to separate it from the air density change, and the air density change, which should originally be able to be learned uniformly, can be learned for each area of the engine operating state. This is because if you have to study and suddenly climb to a high altitude, you will not be able to study each area and your learning will not actually progress.

In view of the above-mentioned problems in the conventional technology, the object of the present invention is to provide an air-fuel ratio learning control device that can quickly and reliably learn changes in air density and can effectively control the air-fuel ratio learning when driving up a mountain, for example.

<Means for solving the problem> In order to achieve the above purpose, the present invention uses a uniform learning correction coefficient mainly for altitude correction to uniformly learn air density changes, and parts. This is a condition where only changes in air density can be learned by dividing the system into area-based learning correction for learning variations in each area of the engine operating state, that is, an area where there is no variation in the system with respect to changes in throttle valve opening. However, in the region where the intake air flow rate hardly changes with respect to the throttle valve opening change at each engine speed (hatched area in Figure 10), the air density change is uniformly learned, and the learning correction coefficient is uniformly adjusted. The configuration is such that the area-specific learning correction coefficient is corrected by learning the component variations and the like in other areas for each area. and again,
In order to give priority to learning about the uniform learning correction coefficient, a large weight is given to the learned value when the uniform learning correction coefficient is corrected, and a small weight is given to the learning value when the area-specific learning correction coefficient is corrected.

Therefore, the air-fuel ratio learning control device according to the present invention is configured to include the following means A to L, as shown in FIG.

(A) Engine operating state detection means that detects the engine operating state including at least parameters related to the amount of air taken into the engine. (B) Detects engine exhaust components and thereby detects the air-fuel ratio of the engine intake mixture. Air-fuel ratio detection means (C) Basic fuel injection amount setting means for setting a basic fuel injection amount based on the parameters detected by the engine operating state detection means (D) Basic fuel injection amount setting means for setting the basic fuel injection amount for all areas of the engine operating state. A rewritable uniform learning correction coefficient storage means (E) that stores a uniform learning correction coefficient for uniformly correcting the amount of fuel. rewritable area-specific learning correction coefficient means (F) area-specific learning correction for searching the area-specific learning correction coefficient of the area of the corresponding engine operating state from the area-specific learning correction coefficient storage means based on the actual engine operating state; Coefficient search means (G) A feedback correction coefficient for comparing the air-fuel ratio detected by the air-fuel ratio detection means with a target air-fuel ratio and correcting the basic fuel injection amount so as to bring the actual air-fuel ratio closer to the target air-fuel ratio. Feedback correction coefficient setting means (H) for increasing or decreasing by a predetermined amount the basic fuel injection amount set by the basic fuel injection amount setting means, the uniform learning correction coefficient stored in the uniform learning correction coefficient storage means, Fuel injection amount calculation means (I) for calculating the fuel injection amount based on the area-specific learning correction coefficient searched by the area-specific learning correction coefficient search means and the feedback correction coefficient set by the feedback correction coefficient setting means; Fuel injection means (J) that injects fuel into the engine on and off in response to a drive pulse signal corresponding to the fuel injection amount calculated by the calculation means.At each engine speed, the intake air flow rate changes approximately depending on the throttle valve opening uniform learning area detection means (K) for detecting a predetermined area that does not change; when the uniform learning area detection means detects that the predetermined area is within the predetermined area, learning the deviation of the feedback correction coefficient from the reference value; , uniform learning correction coefficient correction means for setting a new uniform learning correction coefficient by adding this deviation to the current uniform learning correction coefficient by a predetermined ratio, and rewriting the uniform learning correction coefficient in the uniform learning correction coefficient storage means ( L) When the uniform learning area detecting means does not detect that the area is in the predetermined area, the deviation from the reference value of the feedback correction coefficient is learned for each area of the engine operating state, and the deviation from the reference value is set as the current learning correction coefficient for each area. Area-specific learning correction coefficient correction means for setting a new area-specific learning correction coefficient by adding a predetermined proportion of this deviation and rewriting the area-specific learning correction coefficient in the area-specific learning correction coefficient storage means; M ALT is the addition ratio of the deviation when the uniform learning correction coefficient is modified by the correction coefficient modification means K, and M _ALT is the addition ratio of the deviation when the area-specific learning correction coefficient is modified by the area-based learning correction coefficient modification means L. M _MAP
Then, set the relationship M _ALT > M _MAP .

<Operation> The basic fuel injection amount setting means C sets the basic fuel injection amount corresponding to the target air-fuel ratio based on parameters related to the amount of air taken into the engine, and the area-based learning correction coefficient search means F: The area-specific learning correction coefficient search means E searches for the area-specific learning correction coefficient of the area corresponding to the actual engine operating state, and the feedback correction coefficient setting means G compares the actual air-fuel ratio with the target air-fuel ratio and calculates the actual The feedback correction coefficient is set by increasing or decreasing a predetermined amount based on, for example, proportional/integral control so that the air-fuel ratio of the target air-fuel ratio approaches the target air-fuel ratio. Then, the fuel injection amount calculation means H is
The fuel injection amount is calculated by correcting the basic fuel injection amount using the uniform learning correction coefficient stored in the uniform learning correction coefficient storage means D, correcting it using the area-specific learning correction coefficient, and further correcting it using the feedback correction coefficient. do. Then, the fuel injection means I is actuated by a drive pulse signal corresponding to this fuel injection amount to inject and supply fuel to the engine.

On the other hand, the uniform learning area detecting means J detects whether or not the intake air flow rate does not change with respect to the change in opening of the throttle valve at each engine speed. , the uniform learning correction coefficient storage means K learns the deviation of the feedback correction coefficient from the reference value, and sets a new uniform learning correction coefficient by adding this deviation to the current uniform learning correction coefficient by a predetermined ratio. Then, the data in the learning correction coefficient storage means D is uniformly rewritten. In this way, under the condition that only the change in air density can be learned, that is, in the region where the intake air flow rate almost does not change with respect to the change in opening of the throttle valve at each engine speed (hatched area in Fig. 10), the change in air density can be learned. uniformly learn. It should be noted that this region is not without component variations, but this is a high opening range of the throttle valve, and compared to a low opening range, the main component variation is the pulse width of the fuel injection valve - injection. Variations in flow rate characteristics and intake air amount characteristics relative to throttle valve opening are extremely small, and can be absorbed and learned based on air density.

If the area is outside the predetermined area, the area-specific learning correction coefficient correction means L learns the deviation from the reference value of the feedback correction coefficient for each area of the engine operating state, and incorporates this deviation into the current area-specific learning correction coefficient. A new learning correction coefficient for each area is set by adding a predetermined ratio, and the data in the learning correction coefficient for each area E is rewritten. In this way, parts variations and the like are learned for each area.

Here, the addition ratio M _ALT of the deviation when correcting the uniform learning correction coefficient in the uniform learning correction coefficient correction means K is set to Addition ratio of the above deviation
Since it is larger than the M _MAP , it is possible to perform area-specific learning after uniformly learning about changes in air density with priority.

<Example> An embodiment of the present invention will be described below.

In FIG. 2, the engine 1 includes an air cleaner 2, a throttle body 3, and an intake manifold 4.
Air is inhaled through.

A throttle valve 5 that operates in conjunction with an accelerator pedal (not shown) is provided within the throttle body 3, and a fuel injection valve 6 serving as fuel injection means is provided upstream of the throttle valve 5. The fuel injection valve 6 is an electromagnetic fuel injection valve that opens when a solenoid is energized, and closes when the energization is stopped.The fuel injection valve 6 opens when energized by a drive pulse signal from a control unit 14 (not shown), which will be described later. Fuel is injected and supplied from the fuel pump and adjusted to a predetermined pressure by the pressure regulator. Although this example is a single-point injection system, it may be a multi-point injection system in which a fuel injection valve is provided for each cylinder in a branch of an intake manifold or in an intake port of an engine.

An ignition plug 7 is provided in the combustion chamber of the engine 1. A high voltage generated by an ignition coil 8 is applied to the ignition plug 7 via a distributor 9 based on an ignition signal from a control unit 14, thereby igniting a spark to ignite and burn the air-fuel mixture.

Exhaust gas is discharged from the engine 1 via an exhaust manifold 10, an exhaust duct 11, a three-way catalyst 12, and a muffler 13.

The control unit 14 includes a CPU, ROM,
Equipped with a microcomputer that includes RAM, an A/D converter, and an input/output interface, it receives input signals from various sensors, performs arithmetic processing as described below, and controls the fuel injection valve 6 and ignition coil 8. Control operation.

As the various sensors mentioned above, a potentiometer-type throttle sensor 15 is provided on the throttle valve 5, and outputs a voltage signal corresponding to the opening degree α of the throttle valve 5. Also provided within the throttle sensor 15 is an idle switch 16 that is turned on when the throttle valve 5 is in the fully closed position.

Further, a crank angle sensor 17 is provided built into the distributor 9 to detect the crank angle.
Position signals every 2 degrees and every 180 degrees of crank angle (4
outputs a reference signal (in the case of a cylinder).
Here, the engine rotation speed N can be calculated by measuring the number of pulses of the position signal or the cycle of the reference signal per unit time.

Further, a water temperature sensor 18 that detects the engine cooling water temperature Tw, a vehicle speed sensor 19 that detects the vehicle speed VSP, and the like are provided.

These throttle sensor 15, crank angle sensor 17, etc. are means for detecting the engine operating state.

In addition, an O ₂ sensor 20 is installed in the exhaust manifold 10.
is provided. This O ₂ sensor 20 is a known sensor whose electromotive force suddenly changes when the air-fuel mixture is combusted near the stoichiometric air-fuel ratio, which is the target air-fuel ratio. Therefore, the O ₂ sensor 20 detects the air-fuel ratio (rich).
lean) detection means.

Furthermore, the control unit 14 is equipped with a battery 2 as its operating power source and for detecting current and voltage.
1 is connected via an engine key switch 22. In addition, in the control unit 14
As the operating power source for the RAM, a battery 21 is connected via a suitable stabilized power source without using the engine key switch 22 in order to retain the memory contents even after the engine key switch 22 is turned off.

Here, the CPU of the microcomputer built in the control unit 14 is
Programs on the ROM shown as flowcharts in Figures to Figure 9 (fuel injection amount calculation routine, feedback control zone determination routine, proportional/integral control routine, learning routine, K _ALT learning subroutine, K _MAP learning subroutine, initialization routine) ) to control fuel injection.

In addition, as a basic fuel injection amount setting means, an area-based learning correction coefficient search means, a feedback correction coefficient setting means, a fuel injection amount calculation means, a uniform learning area detection means, a uniform learning correction coefficient correction means, and an area-specific learning correction coefficient correction means. The functions are achieved by the program. Furthermore, RAM is used as the uniform learning correction coefficient storage means and the area-specific learning correction coefficient storage means.

Next, the arithmetic processing of the microcomputer in the control unit 14 will be explained with reference to the flowcharts shown in FIGS. 3 to 9.

In the combustion injection amount calculation routine shown in Fig. 3, Step 1 (marked as S1 in the figure; the same applies hereinafter)
Then, the throttle valve opening α detected based on the signal from the throttle sensor 15 and the engine speed N calculated based on the signal from the crank angle sensor 17 are read.

In step 2, the amount of intake air Q corresponding to the throttle valve opening α and the engine speed N is determined in advance through experiments, etc., and is referred to a stored map in the ROM to search for Q corresponding to the actual α and N. and load it.

In step 3, the intake air flow rate Q and engine speed N
The basic fuel injection amount Tp=K·Q/N (K is a constant) corresponding to the intake air amount per unit rotation is calculated from. Here, steps 1 to 3 correspond to basic fuel injection amount setting means.

In step 4, the rate of change in the throttle valve opening α detected based on the signal from the throttle sensor 15 or from the ON of the idle switch 16 is determined.
Various correction coefficients COEF are set, including an acceleration correction due to switching to OFF, a water temperature correction corresponding to the engine cooling water temperature Tw detected based on the signal from the water temperature sensor 18, and the like.

In step 5, the uniform learning correction coefficient _KALT stored at a predetermined address in the RAM serving as uniform learning correction coefficient storage means is read. Note that the uniform learning correction coefficient _KALT is stored as an initial value of 0 at the time when learning has not started, and this is read.

In step 6, the area-specific learning correction coefficient K _MAP is stored in the RAM as an area-specific learning correction coefficient storage means corresponding to the engine speed N representing the engine operating state and the basic fuel injection amount (load) Tp. Refer to the map, search for and read the K _MAP corresponding to the actual N and Tp. This part corresponds to the area-by-area learning correction coefficient search means. The area-specific learning correction coefficient K _MAP map divides the areas of the engine operating state into an 8x8 grid with the horizontal axis representing the engine speed N and the vertical axis representing the basic fuel injection amount Tp. A separate learning correction coefficient K _MAP is stored,
All initial values are 0 when learning has not started.
has been memorized.

In step 7, the feedback correction coefficient LAMBDA set by the proportional/integral control routine shown in FIG. 5, which will be described later, is read. Note that the reference value of this feedback correction coefficient LAMBDA is 1.

In step 8, a voltage correction amount Ts is set based on the voltage value of the battery 21. This is to correct changes in the injection flow rate of the fuel injection valve due to changes in battery voltage.

In step 9, the fuel injection amount Ti is calculated according to the following equation. This part corresponds to the fuel injection amount calculation means Ti=TpCOEF・(LAMBDA+K _ALT +K _MAP )+
Ts In step 10, the calculated Ti is set in the output register. As a result, at a predetermined fuel injection timing synchronized with engine rotation (for example, every 1/2 rotation), a drive pulse signal having a pulse width of Ti is applied to the fuel injection valve 6, and fuel injection is performed.

Figure 4 shows the feedback control zone determination routine, which in principle performs air-fuel ratio feedback control at low to medium speeds and low to medium loads, and stops air-fuel ratio feedback control at high speeds or high loads. It is.

In step 21, a comparison Tp is searched from the engine speed N, and in step 22, the actual basic fuel injection amount is determined.
Compare Tp (actual Tp) and comparison Tp.

If actual Tp≦comparison Tp, that is, in the case of low-medium speed and low-medium load, proceed to step 23 to reset the delay timer (which is counted up by the clock signal), and then proceed to step 26 to set the λ _CONT flag. Set to 1. This is to perform air-fuel ratio feedback control in the case of low to medium speeds and low to medium loads.

If actual Tp>comparison Tp, that is, if the rotation is high or the load is high, the process proceeds to step 27 and the λ _CONT flag is set to 0, in principle. This is to stop the air-fuel ratio feedback control, separately obtain a rich output air-fuel ratio, suppress the rise in exhaust temperature, and prevent the engine 1 from seizing and the catalyst 12 from burning out.

Here, even in the case of high rotation or high load,
By comparing the value of the delay timer with a predetermined value in step 24, after shifting to high rotation or high load, the process proceeds to step 26 and continues to set the λ _CONT flag to 1 until a predetermined time has elapsed, and the air-fuel ratio is Continue feedback control. This is to increase the opportunity to learn about the uniform learning correction coefficient K _ALT since mountain climbing is performed in a high rotation/high load range. However, if the engine speed N exceeds a predetermined value (for example, 3800 rpm) as determined in step 25, the air-fuel ratio feedback control is stopped for safety reasons.

FIG. 5 shows a proportional/integral control routine, which is executed at predetermined intervals (for example, 10 ms), thereby setting the feedback correction coefficient LAMBDA. Therefore, this routine corresponds to feedback correction coefficient setting means.

In step 31, the value of the λ _CONT flag is determined,
If it is 0, this routine ends. In this case, the feedback correction coefficient LAMBDA is clamped to the previous value (or reference value 1), and the feedback control of the air-fuel ratio is stopped.

If the λ _CONT flag is 1, the process advances to step 32 to read the output voltage V ₀₂ of the O ₂ sensor 20, and in the next step 33 the air-fuel ratio is enriched by comparing it with the slice level voltage V _ref corresponding to the stoichiometric air-fuel ratio.・
Determine the zone.

When the air-fuel ratio is lean (V ₀₂ <V _ref ), the process advances from step 33 to step 34 to determine whether or not it is the time of reversal from rich to lean (immediately after reversal), and when the ratio is reversed, the process advances to step 35. The feedback correction coefficient LAMBDA is increased by a predetermined proportionality constant P relative to the previous value. Step 3 except when reversing
6, the feedback correction coefficient LAMBDA is increased by a predetermined integral constant I with respect to the previous value, and thus the feedback correction coefficient LAMBDA is increased at a constant slope. Note that P≫I.

When the air-fuel ratio is rich (V ₀₂ >V _ref ), the process proceeds from step 33 to step 37 to determine whether or not it is the time of reversal from lean to rich (immediately after reversal), and when the air-fuel ratio is reversed, the process proceeds to step 38. The feedback correction coefficient LAMBDA is decreased by a predetermined proportionality constant P from the previous value. Step 3 except when reversing
9, the feedback correction coefficient LAMBDA is decreased by a predetermined integral constant I from the previous value, and thus the feedback correction coefficient LAMBDA is decreased at a constant slope.

FIG. 6 shows a learning routine, FIG. 7 shows a K _ALT learning subroutine, and FIG. 8 shows a K _MAP learning subroutine.

At step 41 in FIG. 6, the value of the λ _CONT flag is determined. If it is 0, the routine proceeds to step 42, where the count values C _ALT and C _MAP are cleared, and this routine ends. This is because learning cannot be performed when air-fuel ratio feedback control is stopped.

When the λ _CONT flag is 1, that is, during air-fuel ratio feedback control, the process proceeds to step 43 and subsequent steps to learn the uniform learning correction coefficient _KALT (hereinafter referred to as ``1'').
K _ALT learning) and area-specific learning correction coefficient K _MAP
(hereinafter referred to as K _MAP learning).

That is, K _ALT learning is performed in a prescribed high load region (hereinafter referred to as the Q flat region) where the intake air flow rate Q changes little with respect to changes in the throttle valve opening α at each engine speed N as shown by hatching in Fig. 10, while K _MAP learning is performed in other regions, so that in step 43 a comparison _α1 is retrieved from the engine speed N, and in step 44 the actual throttle valve opening α (actual α) is compared with comparison _α1 . These steps 43 and 44 correspond to uniform learning region detecting means.

As a result of comparison, actual α≧comparison α ₁ (Q flat region)
In this case, in principle, proceed to steps 48 and 49, clear the count value C _MAP , and then proceed to steps 48 and 49 as shown in Figure 7.
K Executes _{the ALT} learning subroutine.

However, in the case of a single-point injection system, in the region where the throttle valve opening is extremely large, the intake flow velocity becomes slow and the distribution to each cylinder deteriorates. Assign this to prohibit K _ALT learning at throttle valve openings greater than that. Therefore, in step 45, a comparison α ₂ is searched from the engine speed N, and in a step 46, the actual α and the comparison _{α 2 are} compared.
After clearing the count value C _ALT , move to the 8th
Move to the K _MAP learning subroutine shown in the figure.

In addition, in the case of a single point injection system, the distance from the fuel injection valve 6 to the combustion chamber of the engine 1 is long, and during sudden acceleration, the wall flow of fuel causes
Accurate K _ALT learning cannot be performed, so when accelerating suddenly, wait for the set time, that is, until the wall flow becomes steady.
K _ALT study. Therefore, it is determined in step 47 whether a predetermined time has elapsed after acceleration, and if it has not elapsed, the process proceeds to steps 50 and 51, and after clearing the count value _CALT , the K _MAP learning shown in Fig. 8 is performed. Move to subroutine.

If the judgment in step 44 is that actual α<comparison α ₁ , proceed to steps 50 and 51 and calculate the count value.
After clearing C _ALT , move to the K _MAP learning subroutine shown in Figure 8.

Next, the _KALT learning subroutine shown in FIG. 7 will be explained. This _KALT learning subroutine corresponds to a uniform learning correction coefficient correction means.

In step 61, it is determined whether the output of the O ₂ sensor 20 has been reversed, that is, the direction of increase/decrease of the feedback correction coefficient LAMBDA has been reversed, and when this subroutine is repeated to reverse the output, a count value _CALT representing the number of reversals is determined in step 62. 1 and reaches, for example, C _ALT = 3, the process proceeds from step 63 to step 64 and calculates the deviation of the current feedback correction coefficient LAMBDA from the reference value 1 (LAMBDA
-1) is temporarily stored as ΔLAMBDA ₁ , and learning is started.

Then, when C _ALT = 4 or more, step 63
The process then proceeds to step 65, where the deviation (LAMBDA-1) of the feedback correction coefficient LAMBDA from the reference value 1 at that time is temporarily stored as _ΔLAMBDA2 . ΔLAMBDA ₁ and ΔLAMBDA ₂ stored at this time are the deviations from the reference value 1 of the feedback correction coefficient LAMBDA from the previous (e.g., third) reversal to the current (e.g., fourth) reversal, as shown in FIG. These are the upper and lower peak values of .

In this way, the feedback correction coefficient
Upper and lower peak values of LAMBDA deviation from standard value 1 ΔLAMBDA ₁ . Once ΔLAMBDA ₂ is determined, the process proceeds to step 66, where their average value (see the following equation) is determined.

= (ΔLAMBDA ₁ +
ΔLAMBDA ₂ )/2 Next, the process proceeds to step 67 where the current uniform learning correction coefficient stored at a predetermined address in the RAM is calculated.
Read K _ALT (initial value 0).

Next, the process proceeds to step 68, and a new uniform learning correction coefficient is created by adding a predetermined percentage of the average value of the deviation from the reference value 1 of the feedback correction coefficient to the current uniform learning correction coefficient _KALT according to the following formula. Calculate K _ALT and correct and rewrite the uniform learning correction coefficient data at a predetermined address in RAM.

K _ALT ← K _ALT + M _ALT (M _ALT is an addition ratio constant, 0<M _ALT <1) After this, in step 69, ΔLAMBDA ₂ is substituted into ΔLAMBDA ₁ for the next learning.

Then, in step 70, the learning counter is incremented by 1 using _KALT . The _KALT learning counter has been set to 0 by the initialization routine shown in FIG. 9, which is executed when the engine key switch 22 (or start switch) is turned on.
K _ALT after turning on engine key switch 22
Counting the number of times you learn.

Next, the K _MAP learning subroutine shown in FIG. 8 will be explained. This K _MAP learning subroutine corresponds to area-specific learning correction coefficient correction means.

In step 81, it is determined whether the engine speed N and basic fuel injection amount Tp, which indicate the engine operating condition, are in the same area as the previous time. If the area has changed, the process proceeds to step 82, where the count value _CMAP is cleared, and this subroutine is then terminated.

If it is the same area as last time, go to step 83.
It is determined whether the output of the O ₂ sensor 20 has been reversed, that is, the direction of increase or decrease of the feedback correction coefficient LAMBDA has been reversed, and each time this subroutine is repeated and the output is reversed, the count value C _MAP representing the number of times of reversal is incremented by 1 in step 84. For example, when C _MAP = 3, the process proceeds from step 85 to step 86, where the deviation (LAMBDA - 1) of the current feedback correction coefficient LAMBDA from the reference value 1 is temporarily stored as ΔLAMBDA ₁ , and learning is started. do.

Then, when C _MAP = 4 or more, step 85
The process then proceeds to step 87, where the deviation (LAMBDA-1) of the feedback correction coefficient LAMBDA from the reference value 1 at that time is temporarily stored as ΔLAMBDA ₂ .

In this way, the feedback correction coefficient
Once the upper and lower peak values ΔLAMBDA ₁ and ΔLAMBDA ₂ of deviation from the reference value 1 of LAMBDA are determined, the process proceeds to step 88 and their average value is determined.

Next, the process proceeds to step 89, where the area-specific learning correction coefficient K _MAP (initial value 0) stored in the map on the RAM corresponding to the current area is retrieved and read out.

Next, the process proceeds to step 90, where the value of the _KALT learning counter is compared with a predetermined value, and if it is less than the predetermined value, an addition ratio constant (weighting constant) _MMAP is set in step 91.
is set to a relatively small value M ₀ including 0. If the value is greater than the predetermined value, the addition ratio constant (weighting constant) M _MAP is set to a relatively large value M ₁ in step 92.
(However, M ₁ ≫ M _ALT ).

Next, the process proceeds to step 93, where a new area-specific learning correction is performed by adding a predetermined percentage of the average value of the deviation from the reference value of the feedback correction coefficient to the current area-specific learning correction _coefficient K MAP according to the following formula. Calculate the coefficient K _MAP and correct and rewrite the area-specific learning correction coefficient data for the same area of the map on the RAM.

K _MAP ←K _MAP +M _MAP・After this, in step 94, ΔLAMBDA ₂ is substituted into ΔLAMBDA ₁ for the next learning.

Regarding the addition ratio constant (weighting constant) mentioned above, M _ALT 〓 M _MAP is set after M _ALT learning related to air density changes is performed first, and then area-specific
This is to make K _MAP learning possible.

Also, depending on the number of K _ALT learnings after the engine key switch 22 (or start switch) is turned on,
Changing the value of M _MAP suppresses the progress of K _MAP learning until K _ALT learning is experienced, and in extreme cases
This is to set K _MAP = 0 and prohibit K _MAP learning.

As a result, if you drive uphill to a high altitude without entering the Q flat region, the K _ALT
K _MAP learning progresses only in some areas, including the change in air density, without sufficient learning progressing.
It is possible to prevent deterioration of drivability, exhaust performance, etc. due to large differences in area-specific learning correction coefficients between areas.

<Effects of the invention> As explained above, according to the invention, since air density changes are learned with uniform priority in the Q flat region, it is possible to learn air density changes at high speed, and uniform learning correction is possible. Since learning about coefficients is always given priority over learning about area-specific learning correction coefficients, it is possible to prevent air density changes from being learned by area in some areas as much as possible. , reliable learning becomes possible,
The advantage is that good air-fuel ratio learning control is possible even when driving up a mountain.

[Brief explanation of the drawing]

Figure 1 is a functional block diagram showing the configuration of the present invention.
Fig. 2 is a system diagram showing an embodiment of the present invention, Figs. 3 to 9 are flowcharts showing arithmetic processing contents, Fig. 10 is a diagram showing a learning area for uniform learning correction coefficients, Fig. 11 FIG. 2 is a diagram showing how the feedback correction coefficient changes. 1... Engine, 5... Throttle valve, 6... Fuel injection valve, 14... Control unit, 15...
... Throttle sensor, 17... Crank angle sensor, 20... O ₂ sensor.

Claims (1)

[Scope of Claim for Utility Model Registration] Engine operating state detection means for detecting the engine operating state including at least parameters related to the amount of air taken into the engine; air-fuel ratio detection means for detecting a fuel ratio; basic fuel injection amount setting means for setting a basic fuel injection amount based on the parameter detected by the engine operating state detection means; A rewritable uniform learning correction coefficient storage means storing a uniform learning correction coefficient for uniformly correcting the injection amount, and an area-specific learning correction coefficient for correcting the basic fuel injection amount for each area of the engine operating state. Area-specific learning that searches the area-specific learning correction coefficient of the area of the corresponding engine operating state from the stored rewritable area-specific learning correction coefficient storage means and the area-specific learning correction coefficient storage means based on the actual engine operating state. a correction coefficient search means; and a feedback correction coefficient for comparing the air-fuel ratio detected by the air-fuel ratio detection means with a target air-fuel ratio and correcting the basic fuel injection amount so as to bring the actual air-fuel ratio closer to the target air-fuel ratio. a feedback correction coefficient setting means for increasing or decreasing by a predetermined amount the basic fuel injection amount set by the basic fuel injection amount setting means, the uniform learning correction coefficient stored in the uniform learning correction coefficient storage means, and the area. a fuel injection amount calculation means for calculating a fuel injection amount based on the area-specific learning correction coefficient searched by the separate learning correction coefficient search means and the feedback correction coefficient set by the feedback correction coefficient setting means; and the fuel injection amount calculation means A fuel injection means that injects fuel into the engine on and off in response to a drive pulse signal corresponding to the fuel injection amount calculated by , and an intake air flow rate that almost does not change with respect to changes in the opening of the throttle valve at each engine speed. uniform learning area detection means for detecting a predetermined area; and when the uniform learning area detection means detects the predetermined area, learning the deviation of the feedback correction coefficient from a reference value; uniform learning correction coefficient correction means for setting a new uniform learning correction coefficient by adding this deviation to the learning correction coefficient by a predetermined ratio, and rewriting the uniform learning correction coefficient in the uniform learning correction coefficient storage means; When the learning area detecting means does not detect that the area is in the predetermined area, the deviation from the reference value of the feedback correction coefficient is learned for each area of the engine operating state, and this deviation is set as the current learning correction coefficient for each area. area-specific learning correction coefficient correction means for setting a new area-specific learning correction coefficient by adding a proportion of , and correcting and rewriting the area-specific learning correction coefficient in the area-specific learning correction coefficient storage means, and When the addition ratio in the uniform learning correction coefficient correction means is M _ALT and the addition ratio in the area-based learning correction coefficient correction means is M _MAP ,
An air-fuel ratio learning control device for an internal combustion engine, characterized in that M _ALT > M _MAP .