JPS6073021A - Air-fuel ratio control for internal-combustion engine - Google Patents

Abstract

PURPOSE:To improve fuel consumption by spreading the lean control range by controlling air-fuel ratio by using an air-fuel ratio feedback correction coefficient and learned value and controlling the air-fuel ratio to lean side under a prescribed condition. CONSTITUTION:When it is judged that the pressure PM of an intake pipe is within a learning range in case of air-fuel ratio control by a control circuit 61, the learned value is learned. In other words, it is judged that air-fuel ratio is feedback-controlled or not on the basis of the output signal of an O2 sensor, and when the air-fuel ratio feedback correction coefficient FAF is skipped, learning is performed. Lean control is performed so that the air-fuel ratio of the mixed gas is close to the lean side in comparison with an aimed air-fuel ratio when the average value of the air-fuel ratio feedback correction coefficient is within a prescribed range. Therefore, the fuel consumption can be improved by spreading the lean control range.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an air-fuel ratio control method for an internal combustion engine, and in particular, to control the air-fuel ratio to a target air-fuel ratio using an air-fuel ratio feedback correction coefficient and a learned value, and to control the air-fuel ratio under predetermined conditions. The present invention relates to an air-fuel ratio control method for an internal combustion engine that controls the air-fuel ratio to be leaner than a target air-fuel ratio.

[Prior art]

Conventionally, a three-way catalyst is used to simultaneously purify carbon monoxide, hydrocarbons, and nitrogen oxides in exhaust gas, and in order to improve the purification rate of this three-way catalyst, the 02 sensor Feedback control is performed in which the air-fuel ratio is estimated by detecting the residual oxygen concentration in the air, and the air-fuel ratio is controlled to be close to the stoichiometric air-fuel ratio. In performing this feedback control, the output from the 02 sensor is determined during the basic fuel injection time 1゛P determined by the engine load (intake pipe pressure PM or intake air amount Q/NE per engine revolution) and the engine speed. Based on the signal-processed air-fuel ratio signal, the fuel injection time is multiplied by the air-fuel ratio feedback correction coefficient P A F shown in FIG. By opening the fuel injection valve for a time corresponding to the time TAU, the air-fuel ratio is controlled to be close to the stoichiometric air-fuel ratio. However, due to environmental changes and changes over time, changes in valve timing due to changes in tabet clearance, changes in the characteristics of the pressure sensor and air 70-meter, and changes in the characteristics of the fuel injector occur, causing the fuel injection amount to be lower than the engine's required fuel consumption. In some cases, the injection amount cannot be controlled and the air-fuel ratio cannot be controlled close to the stoichiometric air-fuel ratio. For this reason, air-fuel ratio learning control is performed to control the air-fuel ratio so that it is always near the stoichiometric air-fuel ratio.

In recent years, from the perspective of reducing fuel consumption, under certain conditions during feedback control, that is, at light loads after complete warm-up,
It has been proposed to perform feedforward control or one-in control of the air-fuel ratio to leaner than the stoichiometric air-fuel ratio.

The above air-fuel ratio learning control and lean control are as follows (1
) The fuel injection time TAU is calculated based on the formula and a predetermined amount of fuel is injected.

TAU=(TP-+-TAUG)(1+KG)・F'A
F'・FLEAN-F(t)・・・・・・・・・(1)
However, TAUG is the learned value when the intake throttle valve is in the idle position, KG is the learned value when the intake throttle valve is not in the idle position, F(t) is a correction coefficient such as the warm-up increase coefficient and the start-up increase coefficient, FLEA and N are lean correction coefficients of 1,000 yen. In addition, the learned value KG is determined by the engine load. For example, when the intake pipe pressure is between 200 and 300 Hp, the KG is determined by the engine load.
C), KG2.400 when 300-400tmHy
KG is adopted when ~500 rmI-1y.

These learning values TAUG, KO (KG, , JgG2, K
G3) is learned during air-fuel ratio feedback control and when the engine cooling water temperature exceeds a predetermined value (for example, 80° C.). First, every time the air-fuel ratio feedback correction coefficient FAF skips, the arithmetic average value of the maximum and minimum values of the correction coefficient FAF is 1.
=' A I" AV, Susuwachi, 2.2, 2
, the average FAFAVO value is within a specified range (for example, ±2
%) When the value outside the range is reached, the learned value is set by learning ']”AU
Increase or decrease G and KO by a predetermined value. That is, the average value FAI・
When 'AV exceeds 102, the learned values TAUG and KG are increased by a predetermined value, and when the average values FA and FAV become less than 0.98, the learned values TAUG and KG are subtracted by a predetermined value.

Then, the learned values TAUG, KG learned as above
is determined by the above (1
) is applied to the equation, and the fuel injection time TAU is determined. As a result, when the average value 1 ("A F A The air-fuel ratio is controlled to the lean side, and learning control is performed so that the average value F A F A V approaches 1, that is, the stoichiometric air-fuel ratio.

In addition, the lean correction coefficient F L E A N is set to a value of 1 or less based on the intake pipe pressure (or the amount of intake air per engine revolution) or the intake pipe pressure and the engine speed. The IJ-on correction coefficient FLE is adjusted according to the operating condition.
By applying AN to equation (1) above, the fuel injection amount is reduced and lean control is performed. In addition, when the lean control execution conditions are not satisfied, the lean correction coefficient F
Lean control is stopped by setting the L EA KO value to 1.

The top distribution control is executed when a predetermined condition is satisfied during feedback control, but conventionally, in order to change the learned value by learning, it is executed for a predetermined period of time after the predetermined condition is satisfied (for example, :l1wj*), the control shifted to lean control. For this reason, there is a problem in that during the transition from feedback control to lean control, feedback control is performed even though lean control can be performed, resulting in poor fuel efficiency during this period.

In order to solve the above problem, the inventors of the present invention conducted extensive research and found that once the change in the learned value is completed near the stoichiometric air-fuel ratio, the amount of change in the learned value due to subsequent learning is small. From this, the present inventors have found that in a region where the amount of change in the learning value is small, it is not necessary to provide time for learning and immediate IJ-on control is possible.

[Purpose of the invention]

The present invention has been made based on the above findings as well as solving the above problems, and aims to provide an air-fuel ratio control method for an internal combustion engine that expands the lean control range and further improves fuel efficiency. .

[Structure of the invention]

To achieve the above object, the present invention provides an air-fuel ratio feedback correction coefficient obtained based on an output signal of an oxygen sensor that detects the residual oxygen concentration in exhaust gas, and an average value of the air-fuel ratio feedback correction coefficients that is set to a target air-fuel ratio. The basic fuel injection time determined by the engine load and engine speed is corrected using the learned value that is changed by learning so as to approach the value corresponding to , when the average value of the air-fuel ratio feedback correction coefficient is within a predetermined range including the value corresponding to the target air-fuel ratio, perform lean control so that the air-fuel ratio of the air-fuel mixture becomes leaner than the target air-fuel ratio. It is characterized by As a result, air-fuel ratio lean control is performed when the average value of the air-fuel ratio feedback correction coefficient is within a predetermined range that includes the value corresponding to the target air-fuel ratio, that is, when the amount of change in the learned value due to learning is small. .

〔Effect of the invention〕

Therefore, according to the present invention, the lean control range can be expanded to further improve fuel efficiency.

[Embodiments of the invention]

Embodiments of the present invention will be described in detail below based on the drawings.

FIG. 2 shows a configuration example of an internal combustion engine for an automobile including an electronic fuel injection control device to which the present invention is applied. Air filter 1 is connected to throttle body 5 via inlet pipe 3.

The throttle body 5 is provided with a fuel injection valve 7 on its upstream side, and an intake throttle valve 9 is provided downstream of the fuel injection valve 7 to adjust the amount of intake air in conjunction with an accelerator pedal (not shown). , an intake pipe absolute pressure sensor 11 is provided downstream of the intake throttle valve 9 to measure the absolute pressure at that location. Furthermore, a valve opening position sensor 2 that measures the opening position of the intake throttle valve 9, an idle switch 4 that is turned on only when the intake throttle valve 9 is in the idle position (when it is fully closed), and an intake throttle sensor 2 that measures the opening position of the intake throttle valve 9; A power switch 6 that is turned on only when the opening degree of the valve 9 is 30 degrees or more is attached in association with the intake throttle valve 9.

The throttle body 5 is connected to an intake manifold 13 having branch pipes connected to each cylinder of the engine,
The intake manifold 13 is provided with an intake temperature sensor 15 that measures the temperature of intake air therein. A riser portion 17 is provided on the bottom wall 13a of the intake manifold 13 before branching, through which engine cooling water is circulated to heat the air-fuel mixture.

Reference numeral 19 designates a well-known engine body, which includes a piston 21, a cylinder 23, and a cylinder head 25, and a combustion chamber 27.
The air-fuel mixture sucked into the combustion chamber 27 via the intake valve 29 is ignited by the ignition plug 31 . A water jacket 33 is formed around the cylinder 23, and engine cooling water is circulated through the walk jacket 33 to cool parts including the cylinder 23. An engine coolant temperature sensor 37 is provided on the outer wall of the cylinder block 35 to measure the temperature of the engine coolant in the meter jacket 33.

An exhaust manifold 39 is connected to an exhaust boat (not shown) of the cylinder head 250, and an 02 sensor 41 for measuring the residual oxygen concentration υF in the exhaust gas is provided on the downstream side thereof. The exhaust manifold 39 is connected to the exhaust pipe 45 via a three-way catalytic converter 43. Reference numerals 7 and 47 are connected to the engine body 19, and this is a transmission, which changes the speed of the vehicle depending on the rotational speed of its final output shaft. Several vehicle speed sensors 49 are shown for measuring speed.

In addition, 51 is a key switch, 53 is an igniter, 55
is Dis (Rebune) and Distributor 5
5 is provided with a Ne sensor 57 that outputs an on/off signal at every predetermined crank angle θ1, and the output signal allows the engine rotation speed and the predetermined crank angle position to be known. A G sensor 59 is provided which outputs an on/off signal at each large angle θ2, and cylinder discrimination and top dead center position detection are performed based on the output signal.

Further, 60 indicates a battery.

Control circuit temperature sensor 15, engine cooling water temperature sensor 37.02 sensor 41, which is composed of a microcomputer, etc.
Vehicle speed sensor 49, key switch 51, Ne sensor 57,
G sensor 59 and battery 6°4, intake temperature signal S5,
A water temperature signal S6, an air-fuel ratio signal S7, a vehicle speed signal S8, an ignition signal s9, an engine speed signal S10, a cylinder discrimination signal 811, and a battery voltage signal 814 are input from each sensor. The control circuit 61 also controls the fuel injection valve 7
and an igniter 53, and outputs a fuel injection signal S12 and an ignition signal 813 based on predetermined calculations.

As shown in FIG. 3, the control circuit 61 includes a central processing unit (CPU) 61a that controls various devices, a central processing unit (CPU) 61a in which various numerical values and programs are written in advance, and a one-doon memory (RO).
M) 61b, Random access memory (RAM) 61c in which numerical values and flags for calculation processes can be written in a predetermined area and sized appropriately;
A/D converter (ADC) 61 d1 which converts analog input signals to digital signals Input/output interface (Ilo) 61 e1 where various digital signals are input and various digital signals are output Powered from the auxiliary power supply and stored when the engine is stopped Backup memory (B
U-RAM) 61f, and path lines 61g to which these devices are respectively connected. Above R
The AM is provided with an area for storing the learned value TAUG%KG, and the ROM is pre-stored with the program of the above equation (1), the map of the lean correction coefficient FLEAN shown in Fig. 4, and other constants. ing.

In the above engine, fuel is injected from the fuel injection valve according to the following procedure. On the other hand, the engine rotation speed Ne calculated based on the engine rotation speed signal S1 is read, and the intake pipe pressure P'M is calculated based on the intake pipe pressure signal S4.
Load. Next, based on the engine rotational speed Ne and the intake pipe pressure PM, the basic fuel injection time TP is calculated from the basic fuel injection time map by two-dimensional interpolation.

Next, the fuel injection time TAU is determined based on the above equation (1), and the fuel injection valve is opened for a time corresponding to this fuel injection time TAU to inject fuel. However, in this embodiment, one learned value 'I' AU G and one learned value ``0'' are adopted, and the learned value T AU G is the above-mentioned (1) regardless of the on/off state of the throttle switch. ), and the learned value KG is applied to the above equation (1) for the entire operating range when the throttle switch is turned on. In addition, when lean control conditions are satisfied, lean correction coefficients F L and E A
N is applied to the above equation (1).

Next, a processing routine of an embodiment in which the present invention is applied to the above engine will be described in detail.

FIG. 5 shows the learning routine of this embodiment that is executed every predetermined time (for example, 48 m5ec).
In step 100, it is determined whether the intake throttle valve 9 is in the idle position or not based on the closed state of the idle switch. When the idle switch is off, in step 101 the intake pipe pressure P~1 changes from 200 to 4. OOmH
Is the intake pipe pressure IJ within the range of y?
Determine whether M is within the learning area. This intake pipe pressure range indicates the intake pipe pressure under steady running conditions. 1. Since the deviation of the J#4Fe average value FA F A V is approximately the same at a rate of 1 in all driving ranges, the learning range is defined only in the steady running state. Intake pipe pressure) J
If hi is within the learning area, step 103
The learning value is learned by determining the following learning conditions, and if the intake pipe pressure PM is not within the learning range, the routine proceeds to the next step without learning.

On the other hand, if the idle switch is on, step 102
, the engine speed Ne is a predetermined value (for example, 100
0 rpm) and whether the intake pipe pressure P%1 exceeds a predetermined value (for example, 200 m+Hy). If the judgment in step 102 is affirmative, that is, in the case of a normal idle link, the following learning conditions are judged and learning values are learned in step 10.
If the determination in step 2 is negative, that is, during cranking or idle up, the process proceeds to the next routine without learning.

In step 103, it is determined whether feedback control is being performed so that the air-fuel ratio becomes the stoichiometric air-fuel ratio, which is the target air-fuel ratio, based on the output signal of the 02 sensor. If feedback control is not in progress, for example if lean control is being performed, abnormal learning may occur, so it is necessary to learn. 'HW
It is determined whether or not the temperature exceeds a predetermined value (for example, 80° C.). When the coolant temperature THW is below a predetermined value, learning is not performed because the engine is warming up, and when the coolant temperature THW exceeds a predetermined value, in step 105, the atmospheric pressure PMA exceeds a predetermined pressure (for example, 65 o m + nHf). Determine whether or not there is. This atmospheric pressure PMA can be detected by taking in the intake pipe pressure signal S4 under conditions where the intake pipe pressure and the atmospheric pressure are equal. This condition is that the intake throttle valve is fully open and the engine speed Ne is a predetermined value (
For example, 2000 rpm) or less. Step 1
When the atmospheric pressure PMA is determined to be below a predetermined value in step 05, that is, when driving at high altitudes, no learning is performed, and when it is determined that the atmospheric pressure exceeds the predetermined value, the learning is performed within a predetermined time (for example, 2 sec) in step 106. Vehicle speed change rate ΔSP
It is determined whether the amount of change in D, that is, the rate of change in vehicle speed is less than a predetermined value (for example, 0.7 km/h). Learning is not performed when the rate of change of the vehicle speed is greater than a predetermined value, that is, when the vehicle is accelerating.

If it is less than the predetermined value, the intake temperature 'I'HA is determined in step 107.
It is determined whether the temperature is within a predetermined range (for example, 40°C<THA<90°C). When the intake air temperature THA is outside the predetermined range, that is, at extremely low temperatures and high temperatures, abnormal learning occurs and no learning is performed, and when the temperature is within the predetermined range, the air-fuel ratio feedback correction coefficient F A F is determined in step 108.
It is determined whether or not it has been skipped, and only when it has been skipped, the learning value is learned in step 09.

An example of the learning in step 109 will be explained based on the processing routine shown in FIG. First, in step 110, it is determined whether the air-fuel ratio correction coefficient FAF has skipped a predetermined number of times, and only when it has skipped a predetermined number of times, in step 111, the average value F'A is determined based on the above equation (2).
I” AV is calculated. Here, the average value is calculated after skipping a predetermined number of times.

This is because the change in the air-fuel ratio feedback correction coefficient is unstable immediately after shifting from lean control, which is oven loop control, to feedback control. Therefore, the unstable air-fuel ratio feedback correction coefficient is not used in calculations.

In the next step 112, the average values p h, p A,
Determine whether or not V exceeds 1. If V exceeds 1, that is, if the air-fuel ratio is as high as the stoichiometric air-fuel ratio, determine whether or not the idle switch is on in step 13.
When the idle switch is on, the learned value TAUG is increased by a predetermined amount (for example, 0.002) in step 114;
When the idle switch is off, the learned value KO is increased by a predetermined amount (for example, 8) in step 115. On the other hand, when the average value T="A In step 118, the learned value TA UG is decreased by a predetermined amount (for example, 0.002), and when the idle switch is off, the learned value K G is decreased by a predetermined amount (for example, 8).

The learning value learned as described above is applied to the equation (1) above according to the on/off state of the throttle switch, and the air-fuel ratio is learning-controlled.

Next, the lean correction coefficient 1・LEAN regarding lean control
The arithmetic processing routine will be explained with reference to FIG. In step 120, the mode condition xMOD
Determine whether E holds true. This mode condition is established when the engine is not in the starting state, when the power is not being increased after starting, and when the output is not being increased in 1', and is determined based on the ignition signal S9 and the engine speed signal S10 as to whether or not the engine is in the starting state. When the ignition switch is on and the engine speed is a predetermined value (for example, 500 r
pm ) or below, it is determined that the engine is in a starting state.

In addition, whether or not the amount is being increased after starting is determined by whether or not the amount increasing coefficient after starting H [which is attenuated by a predetermined amount for each fixed number of injections after starting] is 0. It is determined that the dose is not being increased. Whether or not the output is being increased is determined based on a power increase coefficient that is set to a predetermined value (for example, 0.15) when the power switch 6 is on and the cooling water temperature is above a predetermined value (for example, 20 degrees Celsius). When the power increase coefficient is O, it is determined that the power is not being increased.

If the mode condition xMODE is not satisfied, feedback control is performed with the lean correction coefficient I''LEAN set to 1 in step 128, and the mode condition xMODE is
When E is established, it is determined in step 121 whether the starting temperature correction coefficient ADD is 0 or not.

The starting temperature correction coefficients A and DD are used to correct variations in the air-fuel ratio due to the temperature difference between the engine cooling water temperature and the intake manifold wall surface temperature. In the engine shown in Figure 2, the distance from the fuel injector to the combustion chamber is long, and if the intake manifold wall temperature is low, air-fuel ratio correction based on engine cooling water temperature alone is not sufficient, so this coefficient A is particularly important.
DD is required. The starting temperature correction coefficients A and DD are determined by subtracting a predetermined amount from an initial value determined in inverse proportion to the intake air temperature at a predetermined period to attenuate it to O.

When it is determined in step 121 that the starting temperature correction coefficient ADD is not 0, that is, when starting temperature correction is being performed, the process proceeds to step 128, where the starting temperature correction coefficients A, I are determined.
) When D is O, that is, when the starting temperature correction has been completed, it is determined in step 122 whether the engine cooling water temperatures T)] and W are equal to or higher than a predetermined value (for example, 80° C.). When the coolant temperature THW is less than a predetermined value, that is, during warm-up, the process proceeds to step 128, and when the coolant temperature THW is above the predetermined value, that is, after complete warm-up, it is determined in step 23 whether lean control is being performed. Whether lean control is in progress or not, the balance correction coefficient F
This can be determined based on whether LEAN is 1 or not.

When it is determined in step 123 that lean control is being performed, in step 124 the amount of change in the rate of change ΔSPD of the vehicle speed (hereinafter referred to as second derivative value) ΔSPD/2sec within a predetermined time (for example, 2 seconds) is set to a predetermined value (for example, 53Fm/
h) Full judgment is made as to whether or not there is sudden acceleration by determining whether or not the acceleration is below. When the second-order differential value ΔSPD/2sec exceeds a predetermined value, it is determined that there is a sudden acceleration and the process proceeds to step 128, and when it is less than the predetermined value, the process proceeds to step 125. On the other hand, when lean control is not being performed, that is, when feedback control is being performed, the second differential value ΔSPD/2sec in step 126T is a predetermined value smaller than the predetermined value in step 124 (for example, 0.
7 Km/h) or less, it is determined whether or not the acceleration is slow. In step 127, the average value F A F A V of the air-fuel ratio feedback correction coefficient is determined to be a value corresponding to the stoichiometric air-fuel ratio (i.e. 1 ) (for example, 0.99 (1i'AFAV<1.01)
Determine whether or not the value is within the range. If the judgments in step 126 and step 127 are negative, it is judged that one-turn control is impossible, and the process proceeds to step 128. If both judgments are affirmative, it is judged that one-turn control is possible, and the process proceeds to step 125. move on.

As a result of the above, it is determined that lean control is possible even when the variation part of the learned value is small, that is, when the average value 1" A F A V is within a predetermined range that includes the value corresponding to the stoichiometric air-fuel ratio. Lean control becomes possible even in areas that were judged to be uncontrollable.

In step 125, it is determined whether the load is light by determining whether the opening degree TA of the intake throttle valve is less than or equal to a predetermined value (for example, 30 degrees). The opening degree TA of the intake throttle valve can be determined based on whether or not the power switch is turned on. When the opening degree of the intake throttle valve] A exceeds a predetermined value, that is, when the load is high, the process proceeds to step 128, and when the opening degree E' of the intake throttle valve is less than the predetermined value, that is, when the load is light, the process proceeds to step 128. In step 129, it is determined whether the idle switch is off, and it is determined whether the intake throttle valve is open.

When the idle switch is off, in step 130, the lean correction coefficient 1i'LEAN map shown in FIG. Load the correction coefficient F L EAN into register A by interpolation.

In the next step 131, it is determined whether lean control is being performed. If lean control is being performed, then in step 133, the value of register A is set as the lean correction coefficient FLEAN, and lean control is subsequently performed. On the other hand, when lean control is not in progress, the vehicle speed SPD is set to a predetermined value (for example, 10KITI/
h), it is determined whether the running is steady or not, and if it exceeds a predetermined value, step 133
The value of register A is changed to lean correction coefficient F : r, E A
Lean control is performed with N set to N. On the other hand, when the vehicle speed is less than a predetermined value, that is, when the vehicle is started, the lean correction coefficient FLEAN is set to 1 in step 137, and the lean control is stopped.

When the idle switch is on, the average value NAV of the engine speed within a predetermined time is determined in step 134;
In the next step 135, it is determined whether the average value NA■ exceeds a predetermined value B (for example, 600 rpm).

When the average value NAV is less than the predetermined value B, the lean control is stopped in step 137, and when the average value NAV exceeds the predetermined value B, the lean correction coefficient is set to 11"L in step 136.
Lean control is performed by setting E A N to a predetermined value less than 1 (for example, 0.92). In this way, when the idle switch is turned on, the lean corrector pJF INOJ! : A.N.
By performing lean control by setting 17 to a predetermined value, it is possible to prevent idle instability due to hunting caused by intake pipe pressure fluctuations.

According to the above embodiment, since the learning area is one area and the seven learning values are two, it is possible to reduce the storage area of the memory and reduce the number of words of the program. .

In addition, although the engine in which the basic fuel injection time is determined by the intake pipe pressure and the engine speed and uses one fuel injection valve has been described above, the engine to which the present invention is applied is not limited to this, and the engine The present invention can also be applied to an engine in which the basic fuel injection time is determined based on the amount of intake air per revolution and the engine speed, and to an engine in which each cylinder is provided with a fuel injection valve protruding from the intake manifold.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the nitrous-fuel ratio signal and the feedback correction coefficient, FIG. 2 is a block diagram showing a configuration example of the present invention including an engine to which the present invention is applied, and FIG. 3 is a diagram showing the control circuit of FIG. 2. FIG. 4 is a block diagram showing an example; FIG. 4 is a diagram showing a map of linear correction coefficients; FIG. 5 is a flowchart showing a learning routine in an embodiment of the present invention; FIG. FIG. 7 is a flowchart showing the line correction coefficient calculation routine of the embodiment. 7...Fuel injection valve, 9...Intake throttle valve, 11...
Pressure sensor, 15...Intake temperature sensor, 41...02
Sensor, 49...Vehicle speed sensor, 61...Control circuit.

Claims (1)

[Claims] (1) The average value of the air-fuel ratio feedback correction coefficient obtained based on the output signal of the 02 sensor that detects the residual oxygen concentration in exhaust gas and the air-fuel ratio foot check correction coefficient approaches the value corresponding to the target air-fuel ratio. Using the learning value changed by learning, feedback control is performed so that the air-fuel ratio of the air-fuel mixture becomes the target air-fuel ratio by correcting the basic fuel injection time determined by the engine load and engine speed, and the air-fuel ratio feed I (An internal combustion engine that performs lean control so that the air-fuel ratio of the air-fuel mixture becomes leaner than the target air-fuel ratio when the average value of the correction coefficient is within a predetermined range including the value corresponding to the target air-fuel ratio. Air-fuel ratio control method.