JPH05272380A - Air-fuel ratio control device for engine - Google Patents

Abstract

PURPOSE:To improve exhaust performance further by learning to correct an air-fuel ratio feedback control constant by an O2 sensor on the back of a catalyst, and controlling even the amplitude size of a feedback correction coefficient alpha. CONSTITUTION:An air-fuel ratio feedback control constant is set by a setting means based on a signal from an O2 sensor 31 provided at the front of a three way catalyst. Using this control constant, a calculation means 33 calculates an air-fuel ratio feedback correction quantity alpha, and a fuel quantity supplied to an intake tube is controlled by a control means 34 using alpha. In the meanwhile, an air-fuel ratio feedback control constant is learned and corrected by a learning correction means 36 based on a signal from an O2 sensor provided at the back of the three way catalyst. In this case, alpha may overflows out of a window, so the air-fuel ratio feedback control constant is corrected to set the amplitude of alpha to be within a specified width corresponding to the window.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an engine air-fuel ratio control system, and more particularly to a system of a dual O ₂ sensor system.

[0002]

2. Description of the Related Art O ₂ sensors are provided before and after a three-way catalyst,
There is a so-called dual O ₂ sensor system device (“Internal combustion engine”, Vol. 30 No. 381 199).
1.7 page 72, Japanese Patent Application Laid-Open No. 3-217636).

To explain this, in FIG. 11, CO,
The three-way catalyst most efficiently purifies three types of harmful substances, HC and NOx, when the air-fuel ratio of the air-fuel mixture is within a predetermined width (called a window) centered around λ = 1.0. is there. In the three-way catalyst system using the O ₂ sensor, the fact that this sensor can detect the stoichiometric air-fuel ratio point is utilized to control the air-fuel ratio feedback correction coefficient α within the above window to purify the exhaust gas at once. To do.

The purification rate of the three-way catalyst is generally shown as a function of the excess air ratio λ as shown in FIG. 11, and λ = 1.0.
Means the theoretical air-fuel ratio, and when it is smaller than this, rich,
When it is larger than this, it is lean.

The output Vs of the O ₂ sensor provided in front of the three-way catalyst greatly changes around λ = 1.0, as shown in FIG. 11, and is substantially 1 V on the rich side, and on the lean side. Since almost 0 V is output, the slice level Vth is
By setting the output level of the O ₂ sensor (usually around 0.5 V) corresponding to the theoretical air-fuel ratio point (λ = 1.0) and comparing this slice level Vth with the O ₂ sensor output Vs, Vs If> Vth, rich is determined, and if Vs <Vth, lean is determined.

However, it is attached to the exhaust pipe at 900 ° C.
The output characteristics of the O ₂ sensor placed in the exhaust gas vary with respect to the air-fuel ratio due to deterioration over time. 11
The output characteristics of the O ₂ sensor shown in ( ₂₎ shift to the rich side or the lean side, resulting in erroneous output.

When the sensor output deviates to the lean side, the theoretical air-fuel ratio point erroneously detected by the deteriorated sensor is shifted to the lean side by a predetermined value as shown by the broken line in FIG. When the erroneous stoichiometric air-fuel ratio point is set to the control center of α (shown by the broken line at the positions where the upper and lower areas are hatched in the figure are equal), PI control (one of feedback control) is performed, and the lean side is Since the control leans toward the air-fuel ratio, the air-fuel ratio feedback correction coefficient α exceeds the window as shown in FIG. 12, and a large amount of NOx is emitted.

By the way, in the rear of the three-way catalyst, the unburned components in the exhaust gas have completed the reaction by the catalyst as compared with the front, so the O ₂ sensor is a stable equilibrium gas measurement, and an ideal output is obtained. Show. Further, since the clean exhaust gas purified by the catalyst is monitored, the sensor is less poisoned and deteriorated.

Therefore, if an O ₂ sensor is also installed behind the catalyst, the sensor accurately detects the stoichiometric air-fuel ratio point, so in the example of FIG. 12 in which the air-fuel ratio leans toward the lean side, naturally, the lean air-fuel ratio is lean. To be judged.

In this case, in order to return the control center of α to the position of λ = 1.0, the air-fuel ratio feedback control constant (proportional component or integral component) is changed. For example, if the proportional amount P _L is increased and the proportional amount P _R is decreased as shown in FIG. 13, the control center of α can be returned to λ = 1.0, and α does not easily protrude from the window.

In practice, in order to improve the control accuracy, the proportional portion is learned and corrected by the signal from the O ₂ sensor behind the catalyst.
For example, as will be described later, when the air-fuel ratio is reversed to the rich side: α = α− (P _R −P
HOS) When the air-fuel ratio reverses to the lean side: α = α + (P _L + P
In the case of calculating the air-fuel ratio feedback correction coefficient α by HOS), when the lean is judged by the O ₂ sensor behind the catalyst as in the above example, the learning value introduced for the proportional parts P _R , P _L The PHOS is updated so that it increases by a certain value.

In this way, by correcting the deviation of the theoretical air-fuel ratio point of the control O ₂ sensor in front of the catalyst with the output of the O ₂ sensor provided behind the three-way catalyst, the control O ₂ sensor Even if the characteristics change due to deterioration over time,
Exhaust gas purification is maintained well.

[0013]

By the way, in the conventional apparatus, although the control center of α can be returned to λ = 1.0, on the other hand, the amplitude of α may become large and the window may stick out. , I was emitting harmful gas that does not have to be emitted.

In the above example, by increasing P _L and decreasing P _R , the control center of α can be returned to λ = 1.0 as shown in FIG.
By this control, when α exceeds the window as shown in the figure, HC and CO are discharged. of course,
When α exceeds the lower limit value in the figure, NOx is emitted,
When both the upper and lower limit values are exceeded, both are discharged.

That is, in the conventional apparatus, even if α is out of the window, it suffices if the control center of α is returned to the stoichiometric air-fuel ratio point.

Therefore, an object of the present invention is to further improve the exhaust performance by learning and correcting the air-fuel ratio feedback control constant by means of the O ₂ sensor behind the catalyst, while managing the amplitude of α. To do.

[0017]

The first invention, as shown in FIG. 1, is an O ₂ sensor 31 provided in front of a three-way catalyst.
Means 32 for setting an air-fuel ratio feedback control constant (for example, a proportional amount) based on the signal of, and means 3 for calculating the air-fuel ratio feedback correction amount α using this control constant.
3, a means 34 for controlling the amount of fuel supplied to the intake pipe by using this air-fuel ratio feedback correction amount α, and a signal of the air-fuel ratio feedback control constant of an O ₂ sensor 35 provided behind the three-way catalyst. In the air-fuel ratio control device for the engine, which comprises means 36 for learning correction based on
A means 37 for correcting the air-fuel ratio feedback control constant (for example, the control constant itself or a learning value for the control constant) is provided so that the amplitude of the air-fuel ratio feedback correction amount α falls within a predetermined width corresponding to the window.

In a second aspect of the invention, the predetermined width of the first aspect of the invention is widened when the amplitude of the air-fuel ratio feedback correction amount α becomes equal to or less than a predetermined value.

In a third aspect of the invention, the predetermined width of the first aspect of the invention is set for each operating region.

In a fourth aspect of the invention, the predetermined width of the first aspect of the invention is narrowed as the degree of deterioration of the three-way catalyst increases.

According to a fifth aspect of the present invention, the amplitude of the air-fuel ratio feedback correction amount α of the first aspect is the difference between the average value of the upper peaks and the average value of the lower peaks of the air-fuel ratio feedback correction amount α.

[0022]

[Operation] The output fluctuation occurring in the O ₂ sensor in front of the catalyst is α
It appears that the control center of is shifted to rich and lean.

When this deviation is detected by the O ₂ sensor behind the catalyst, the control center of α is the theoretical air-fuel ratio point (λ = 1.0).
The air-fuel ratio feedback control constant is learned and corrected so as to be returned to
The amplitude of becomes large, and α may extend beyond the window.

In this case, if the air-fuel ratio feedback control constant itself and the learning value for the control constant are corrected so that the amplitude of α falls within a predetermined width corresponding to the window, α will not extend beyond the window.

Even while the amplitude of α is within the predetermined width, the control center of α is controlled to the stoichiometric air-fuel ratio point based on the signal from the O ₂ sensor behind the catalyst.

Thus, even when the control center of α is returned to the stoichiometric air-fuel ratio point by the signal of the O ₂ sensor behind the catalyst,
No harmful components are wasted.

In the second aspect of the invention, the predetermined width of the first aspect of the invention is set so that the amplitude of α is equal to or less than a predetermined value, in accordance with the characteristics of the catalyst that purification efficiency becomes higher when the air-fuel ratio fluctuates. When it becomes wide, the purification efficiency of the catalyst will be kept high.

In the third invention, when the predetermined width of the first invention is set for each operating region, the catalyst performance which is different for each operating region is well matched, and harmful components are discharged in all operating regions. Without maximizing the performance of the catalyst.

The width of the window is large when the catalyst is new and becomes small as it deteriorates. Therefore, in the fourth invention,
If the predetermined width of the first aspect of the invention is narrowed as the degree of deterioration of the three-way catalyst advances, the amplitude of α is kept in the window just after the catalyst is deteriorated, just as when the catalyst is new.

On the other hand, the amplitude of α in the first invention is the difference between the maximum value of α and the minimum value of α. The maximum value of α has an upper peak of α, and the minimum value of α has a lower peak of α. If used, the variation of these peak values makes the amplitude of α unstable, and when the first invention corrects the learning value for the air-fuel ratio feedback control constant, the learning value also becomes unstable.

In this case, if the difference between the average value of the upper peaks and the average value of the lower peaks of α in the fifth aspect of the invention is the amplitude of α, the variation in the upper and lower peak values is absorbed, and the amplitude of α becomes The learning value stabilizes.

[0032]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In FIG. 2, 7 is an air flow meter for detecting the intake air amount Qa, and 10 is a crank angle sensor for detecting an engine speed Ne and a reference position of a crank angle. Reference numerals 12A and 12B denote O ₂ sensors provided in front of and behind the three-way catalyst 6, which have a characteristic of abruptly changing at the stoichiometric air-fuel ratio, and are on the rich side or lean side of the air-fuel ratio mixture. It outputs what is called a binary value. These sensors are input to the control unit 21 including a microcomputer together with the throttle sensor 9, the water temperature sensor 11, the knock sensor 13, and the vehicle speed sensor 14.

In the control unit 21, when the operating conditions are such that the exhaust gas regulation is carried out, the air-fuel ratio feedback correction coefficient α is set within a predetermined width provided corresponding to the window.
Control.

Further, in the control unit 21,
O ₂ sensor provided in front of the catalyst (referred to as front O ₂ sensor) 12
Considering that the output characteristic of A changes with respect to the air-fuel ratio, and the control center of α deviates from λ = 1.0 due to this change, an O ₂ sensor provided behind the catalyst (called a rear O ₂ sensor) The air-fuel ratio feedback control constant (proportional here) is learned and corrected based on the signal from 12B.

For these controls in the control unit 21, the flowcharts of FIGS. 3 to 6 are assembled.

FIG. 3 is a basic routine for obtaining the air-fuel ratio feedback correction coefficient α, which is executed in rotation synchronization.
This is in accordance with the fuel injection executed in rotation synchronization.

In step 1, it is checked whether or not the air-fuel ratio feedback control condition (abbreviated as "F / B" in the figure) by the front O ₂ sensor is satisfied, and if it is satisfied, the process proceeds to step 2.

In steps 2 to 4, it is judged by comparing the output of the previous O ₂ sensor with the slice level whether the air-fuel ratio is reversed to rich or lean, or whether rich or lean continues. Immediately after the air-fuel ratio is reversed to the rich side or the lean side according to the determination result of, the proportional map is looked up (step 2,
3, 6, Steps 2, 4, 14), and the map for the integral is looked up while the rich and lean are continued (Steps 2, 3, 10, Steps 2, 4, 18).

Here, the proportional components P _R and P _L and the integral components I _R and I _L
Each map value of 1 is assigned in advance with the basic injection pulse width (engine load equivalent amount) Tp and the engine speed Ne as parameters. This is because the O ₂ sensor 12A can only detect whether it is rich or lean, and does not give information on how much is insufficient.

Regarding the integral, a value obtained by multiplying the map value by the engine load (for example, the fuel injection pulse width Ti) is obtained as the final integrals I _R and I _L (steps 11 and 19). This kind of load correction is necessary for α
Since the amplitude of α becomes large in the operating range where the control cycle of becomes long, the exhaust purification performance of the three-way catalyst may deteriorate,
This is because the amplitude of α is almost constant regardless of the control cycle of α.

Immediately after the air-fuel ratio is inverted, the learning value PHOS1 or PHOS2 for the proportional portion is also looked up (steps 8 and 16). Details of steps 8 and 16 are shown in FIG. 6, in which the rear O ₂ sensor (abbreviated as “rear O ₂ ” in the figure and also in FIG. 4) has not failed (abbreviated as “OK” in the figure. 4 is the same.) Only when you search the map,
The learning value (PHOS1 when reversing to the rich side, PHOS2 when reversing to the lean side) stored in the learning area to which the current driving condition belongs is read (step 7).
1-75).

Here, the learning value P is calculated with respect to the proportional amount P _R (the proportional amount for changing α to the lean side stepwise).
The learning value PHOS2 is prepared separately for H _L from P _L (proportional amount for changing α to the rich side stepwise).

In order to improve the learning accuracy, the learning region is divided into a plurality of learning areas with the basic injection pulse width Tp and the engine speed Ne as parameters, and each learning value PHO
S1 and PHOS2 are stored for each learning area.

In steps 7 and 15 before looking up the learning values PHOS1 and PHOS2, the learning value is corrected (learning correction) using the rear O ₂ sensor output.

FIG. 4 is a subroutine called in steps 7 and 15, and is executed with a cycle in which the output of the previous O ₂ sensor is inverted as a calculation cycle.

In FIG. 4, in steps 31 to 33, it is confirmed whether the following conditions <1> to <3> are satisfied.

<1> After that, the O ₂ sensor has not failed (step 31).

<2> Learning is not prohibited (step 32). For example, learning is prohibited unless the rear O ₂ sensor or the three-way catalyst is in an active state, or if the operator does not stay in the learning area under the same operating conditions for a certain number of times.

<3> Not in the idle state (step 33). This is because the engine warm-up is prioritized in the idle state, and thus the air-fuel ratio feedback control is stopped in the operating range.

When all of the conditions <1> to <3> are satisfied, it is judged that the learning condition is satisfied, and the learning value PHOS stored in the learning area to which the current driving condition belongs.
Both 1 and PHOS2 are looked up and stored in a register in the CPU (step 34).

In step 36, the output of the rear O ₂ sensor is compared with the slice level, and when it is determined that the air-fuel ratio is on the rich side, the two learning values PHOS1 and PHOS2 are reduced by a fixed value DPHOS to obtain the learning value. Is updated and the updated value is stored again in the same learning area (steps 36, 37, 39). Learning value PHOS1, P
When both HOS2 are made small, the proportional component (P _R -PHOS1) becomes large in step 9 of FIG. 3, and the proportional component (P _L + PHOS2) becomes small in step 17, and the control center of α becomes λ = 1.0. It will be returned.

Similarly, when it is judged lean in step 36, the learning values PHOS1 and PHOS2 are
By increasing both by a fixed value DPHOS,
The control center of α can be returned to λ = 1.0 (steps 36, 38, 39).

Here, the learning values PHOS1, PHOS2
Both have been updated, but only one may be used.

Returning to FIG. 3, proportional parts (map values) P _R , P _L
Are corrected by the learning values PHOS1 and PHOS2, and the corrected values (P _R -PHOS1) and (P _L + PHOS) are corrected.
The air-fuel ratio feedback correction coefficient α is calculated using the proportional components of 2) (steps 9 and 17). Integrals I _R , I _L
Also from this, the air-fuel ratio feedback correction coefficient α is calculated (steps 12 and 20).

Immediately after reversing from rich to lean, a proportional amount (P _L + PHOS2) is added to α to return the air-fuel ratio to the rich side with good response, and conversely immediately after reversing from lean to rich, a proportional component from α is obtained. By subtracting (P _R -PHOS1), the air-fuel ratio is returned to the lean side with good response at this time as well. When lean is continuing this time as well, the integral I _L is added to α to slowly return the air-fuel ratio to the rich side, and when rich again this time, the integral I _R is subtracted from α. Also slowly returns the air-fuel ratio to the lean side.

From the thus obtained air-fuel ratio feedback correction coefficient α, a known equation Ti = Tp · Co · α + Ts, where Tp: the basic injection pulse width Co: 1 determined from Ne and Ne and the sum of various increase correction coefficients Ts: battery The fuel injection pulse width Ti given to the injector is calculated by the invalid pulse width corresponding to the voltage. From the injector 4, the fuel amount corresponding to Ti is supplied in synchronization with the engine rotation.

If only the proportional components P _R and P _L as the air-fuel ratio feedback control constants are learned and corrected by using the signal of the rear O ₂ sensor, the amplitude of α becomes large and the window may be pushed out. So, to deal with this,
In this example, steps 35 and 40 are newly added in FIG. Before starting the learning correction from step 36, it is checked in step 35 whether α is within a predetermined width corresponding to the window, and if it is outside, the amplitude of α falls within the predetermined width in step 40. Thus, the proportional portion is corrected.

The predetermined width corresponding to the window is determined by associating the upper limit value MAX and the lower limit value MIN thereof with the lower limit and the upper limit of the window, respectively (since the magnitude relationship between the value of λ and the value of α is opposite). , MAX corresponds to the lower limit of the window, and MIN corresponds to the upper limit of the window). λ = 1.
Since α corresponding to 0 is α = 1.0, the same width on the rich side and the lean side with respect to α = 1.0 (for example, 0.1
), And MAX = 1.1 and MIN = 0.9.

Further, in FIG. 3, the maximum value of α is stored in αmax by moving α to αmax of the memory immediately after the lean to rich inversion (steps 2, 3 and 5), and the inversion from rich to lean is performed. Immediately after doing
By shifting to min, the minimum value of α is stored in αmin (steps 2, 4, 13). An average value of the upper and lower peaks of α (a weighted average value may be used) may be entered in αmax and αmin.

In FIG. 5 (details of step 40), at step 51, the value of αmax and the predetermined upper limit value M are set.
AX is compared, and when the maximum value of α exceeds MAX (when αmax> MAX), the learning value PHOS1 is increased by a fixed value DPHOS and the learning value PHOS2 is set constant in order to reduce the amplitude of α. Only the value DPHOS is reduced (step 52). If the proportional values of (PR-PHOS1) and (PL + PHOS2) are corrected to be smaller by updating the learning value, the amplitude of α becomes smaller than the previous time, and the value of αmax can be set to MAX or less. ..

Similarly, when αmin <MIN in step 56, the value of αmin is increased to MIN or more by decreasing the amplitude of α (step 57).

Further, even when α is out of the predetermined range, the learning values PHOS1 and PHOS2 are updated based on the rear O ₂ sensor output in order to return the control center of α to λ = 1.0 (step 53). 55, steps 58-60).
This control is similar to the case where α does not extend beyond the predetermined range. Following step 52, the rear O ₂ sensor output is compared with the slice level. If the result is rich, only the learned value PHOS1 is used. Is decreased by a fixed value DPHOS, and if lean, it is increased by a fixed value DPHOS (steps 53 to 55). Following step 57, if the comparison result is rich, only the learning value PHOS2 is set to the constant value DP.
HOS is increased, and if lean, DPHOS is decreased by a constant value (steps 58 to 60).

It is also possible to update both learning values PHOS1 and PHOS2 based on the output of the rear O ₂ sensor as in steps 36 to 38.

The operation of this example will now be described.

FIG. 7 shows the waveform of α according to this embodiment when operated under the same conditions as shown in FIG.

As a result of the proportional correction being learned and corrected by the signal of the rear O ₂ sensor, the control center of α is returned to approximately λ = 1.0, and for the time being, the learning values PHOS1 and PHOS2 based on the rear O ₂ sensor are set. We consider that there are no updates.

In this case, if α exceeds the preset upper limit value MAX, immediately after that, the learning value PH
OS1 is increased by a fixed value DPHOS, and learning value PHOS
2 is reduced by DPHOS. As a result, the proportional amount (P _R -PHOS1) given at point A in the figure becomes smaller by DPHOS than the previous time.

Α is a preset lower limit value MIN
Since it does not overflow, the proportional component (P _L +
PHOS2) is also smaller than the previous time by DPHOS. As a result, α does not change from the point C as shown by the broken line, but changes by DPHOS from this point like the solid line below.

That is, if α exceeds MAX even once,
From then on, the amplitude of α is reduced as a whole and returned to within MAX.

In FIG. 7, the amplitude of α is contained between MAX and MIN in one update of the learning value, but when it does not fall within one degree, the update of the learning value is repeated until it falls. Similarly, when α falls below MIN, it is also kept above MIN.

On the other hand, if only the amplitude of α is managed in this way, the amplitude of α can be contained within a predetermined width,
The control center of α may deviate from the theoretical air-fuel ratio point.

In this example as well, as a result of the amplitude of α being contained between MAX and MIN, the control center of α is from H1 to H2 (H1 and H2 are indicated by alternate long and short dash lines) as shown in the central portion of FIG. (Shown) to the lean side.
This is the upper and lower areas S1 and S2 of the hatched figure.
This is because the position where .alpha. Is equal is the control center of .alpha.

However, this shift of the control center is
_The proportional component (P _L + PHOS2) given at point E and the proportional component (P _R −
If both PHOS1) and PHOS1) are decreased by DPHOS from the previous time, the control center of α is returned from H2 to the original H1 (steps 36 and 38 in FIG. 4). After controlling the amplitude of α to be within a predetermined range, the control center of α is returned to the stoichiometric air-fuel ratio point.

In this manner, when the control center of α is returned to the stoichiometric air-fuel ratio point by the signal of the rear O ₂ sensor, by ensuring that α falls within the predetermined width corresponding to the window, harmful components are wasted. So that it is not discharged into the.

By the way, the three-way catalyst has a characteristic that the purification efficiency becomes higher when the air-fuel ratio fluctuates (vertical fluctuation). Therefore, when the amplitude of α becomes smaller than a certain value, the purification efficiency becomes low. Resulting in.

Therefore, a value that the amplitude of α should not become smaller than this is obtained in advance as the minimum regulation value X, and the actual amplitude of α (= αmax-αmin) is measured, and when this becomes smaller than X. Can increase the purification efficiency of the three-way catalyst by increasing the amplitude of α. Therefore, as shown in FIG. 8, the MAX and MIN of the memory are
As an initial value, the upper limit value MAX1 and the lower limit value MIN
1 is set (steps 81 and 82), and the amplitude of α is X.
When it becomes smaller, MAX1 + a (a is a positive constant value) is put in MAX again, and MIN1-a is put in MIN (steps 81 and 83), so that the predetermined width is widened by 2a from the beginning.

FIG. 9 shows a third embodiment. The width of the window in which the three-way catalyst can be efficiently purified differs depending on the operating region due to the influence of temperature, etc. Therefore, if the upper limit value MAX and the lower limit value MIN are the same in all operating regions, the performance of the catalyst can be sufficiently brought out. There may be cases where there is no such substance or harmful components are discharged.

Therefore, as shown in FIG. 9, MAX and MIN are set in accordance with the catalyst performance which is different in each operating region (steps 91 and 92), and in all operating regions,
The performance of the catalyst can be maximized without discharging harmful components. Here, the operation area is divided according to the learning area.

FIG. 10 shows a fourth embodiment. The width of the window is wide when the catalyst is new and narrows as the catalyst deteriorates. Therefore, if the upper limit value MAX and the lower limit value MIN are adjusted when new, after the deterioration progresses, the width between the MAX and MIN , The width of the window becomes narrower, and the amplitude of α exceeds the window. Conversely, MAX and MIN
If they are combined after deterioration, the ability of the catalyst cannot be fully brought out when new.

Therefore, as shown in FIG. 10, the degree of deterioration of the catalyst is diagnosed, and the width between MAX and MIN is narrowed as the degree of deterioration progresses, so that the amplitude of α falls within the window without excess or deficiency. Incidentally, the deterioration diagnosis of the catalyst is known.

The learning values PHOS1 and PHOS are calculated by inserting the average value of the upper peaks of α (or a weighted average value) into αmax and the average value of the lower peaks of α into αmin.
OS2 can be stabilized. Since the values of the upper peak of α and the lower peak of α are different, whether or not the learning condition is satisfied when these values are used (in steps 51 and 56 of FIG. 5, αmax> MAX and αmin <MIN are satisfied. When the learning condition is satisfied), the accuracy of the determination is lowered, but the accuracy of the determination is improved by using each average value of the upper and lower peaks of α, and the learning value is thereby stabilized.

By using the average value of the upper and lower peaks of α for αmax and αmin, the difference α between αmax and αmin can be obtained.
It goes without saying that the amplitude of is stable.

In the embodiment, since the learning value for the proportional portion is updated in order to keep the amplitude of α within a predetermined width, the proportional portion itself can be corrected. In this case, the upper limit value MAX and the lower limit value MIN are selected above and below the theoretical air-fuel ratio point (λ = 1.0) of the rear O ₂ sensor.

Further, in the case where the integrated component is learned and corrected by using the signal of the subsequent O ₂ sensor, the amplitude of α is kept within a predetermined width by updating the learning value for the integrated component or correcting the integrated component itself. You can also

[0085]

According to the first aspect of the present invention, the air-fuel ratio feedback control constant set on the basis of the signal of the O ₂ sensor provided in front of the three-way catalyst is set in the O ₂ sensor provided behind the three-way catalyst. in those learning correction based on the signal, the amplitude of the air-fuel ratio feedback correction amount to correct the air-fuel ratio feedback control constant to fall within a predetermined width in correspondence to the window, the air-fuel ratio feedback by a signal of the rear O ₂ sensor When returning the control center of the correction amount to the stoichiometric air-fuel ratio point, it is possible to prevent wasteful emission of harmful components due to the air-fuel ratio feedback correction amount being out of the window.

In the second aspect of the invention, the predetermined range of the first aspect of the invention is widened when the amplitude of the air-fuel ratio feedback correction amount becomes less than or equal to a predetermined value, so that the purification efficiency of the three-way catalyst is maintained high. it can.

In the third aspect of the invention, the predetermined width of the first aspect of the invention is set for each operating region, so that in all operating regions,
The performance of the catalyst can be maximized without discharging harmful components.

In the fourth aspect of the invention, the predetermined width of the first aspect of the invention is narrowed as the degree of deterioration of the three-way catalyst is increased. Therefore, the amplitude of the air-fuel ratio feedback correction amount can be fit in the window without excess or deficiency.

The fifth aspect of the present invention uses the amplitude of the air-fuel ratio feedback correction amount of the first aspect of the invention as the difference between the average value of the upper peaks and the average value of the lower peaks of the air-fuel ratio feedback correction amount α. The amplitude of the correction amount can be stabilized, and when the first invention corrects the learning value for the air-fuel ratio feedback control constant, the learning value can be stabilized.

[Brief description of drawings]

FIG. 1 is a diagram corresponding to claims of the first invention.

FIG. 2 is a control system diagram of an embodiment.

FIG. 3 is a flowchart for explaining calculation of an air-fuel ratio feedback correction coefficient α.

FIG. 4 is a flowchart for explaining updating of learning values PHOS1 and PHOS2.

FIG. 5 is a flowchart for explaining the upper and lower limits of α and updating of learning values PHOS1 and PHOS2 by the output of the rear O ₂ sensor.

FIG. 6 is a flowchart for explaining lookup of learning values PHOS1 and PHOS2.

FIG. 7 is a change waveform diagram of α in the embodiment.

FIG. 8 is a flowchart of a second embodiment.

FIG. 9 is a flowchart of a third embodiment.

FIG. 10 is a flowchart of a fourth embodiment.

FIG. 11 is a characteristic diagram of a purification rate of a three-way catalyst.

FIG. 12 is a waveform diagram of α in the conventional example.

FIG. 13 is a waveform diagram of α in the conventional example.

[Explanation of symbols]

4 injector 5 exhaust pipe 6 three-way catalyst 7 air flow meter (engine load sensor) 10 crank angle sensor (engine speed sensor) 12A front O ₂ sensor 12B rear O ₂ sensor 21 control unit 31 O ₂ sensor 32 control constant setting means 33 Air-fuel ratio feedback correction amount calculation means 34 Fuel control means 35 O ₂ sensor 36 Learning correction means 37 Control constant correction means

Claims (5)

[Claims] 1. A means for setting an air-fuel ratio feedback control constant based on a signal from an O ₂ sensor provided in front of a three-way catalyst, and a means for calculating an air-fuel ratio feedback correction amount using this control constant. A means for controlling the amount of fuel supplied to the intake pipe by using this air-fuel ratio feedback correction amount, and the air-fuel ratio feedback control constant are learned and corrected based on a signal from an O ₂ sensor provided behind the three-way catalyst. In the engine air-fuel ratio control device including means, means for correcting the air-fuel ratio feedback control constant is provided so that the amplitude of the air-fuel ratio feedback correction amount falls within a predetermined width corresponding to the window. Air-fuel ratio controller for the engine. 2. The air-fuel ratio control device for an engine according to claim 1, wherein the predetermined width is widened when the amplitude of the air-fuel ratio feedback correction amount α becomes equal to or less than a predetermined value. 3. The engine air-fuel ratio control device according to claim 1, wherein the predetermined width is set for each operating region. 4. The air-fuel ratio control device for an engine according to claim 1, wherein the predetermined width is made narrower as the degree of deterioration of the three-way catalyst increases. 5. The amplitude of the air-fuel ratio feedback correction amount is
2. The air-fuel ratio control device for an engine according to claim 1, wherein a difference between an average value of the upper peak and an average value of the lower peak of the air-fuel ratio feedback correction amount is used.