JP4265486B2 - Air-fuel ratio control device for internal combustion engine - Google Patents

Description

  The present invention relates to a characteristic shift detection device for an air-fuel ratio sensor of an internal combustion engine.

  In order to simultaneously purify HC, CO, and NOx in the exhaust gas of an internal combustion engine (engine) with a three-way catalyst, it is necessary to maintain the catalyst atmosphere at a theoretical air-fuel ratio (hereinafter referred to as “stoichiometric”). It is known that the catalyst has an oxygen storage capability (oxygen retention capability) so that the purification efficiency at the time of deviation does not decrease.

  If exhaust that is leaner than stoichiometric is given, the catalyst takes in oxygen in the exhaust, and the catalyst atmosphere can be maintained stoichiometric until this oxygen storage amount is saturated. Further, when exhaust richer than stoichiometric is given, oxygen retained by the catalyst is released, and the catalyst atmosphere is kept stoichiometric until all of the retained oxygen is released. In this way, the catalyst compensates for the excess or deficiency of oxygen resulting from a temporary air-fuel ratio shift, and the catalyst atmosphere can be kept substantially stoichiometric.

  If the air-fuel ratio is controlled so that the oxygen storage amount of the catalyst is always at the target value, for example, about half of the maximum storage amount, the oxygen intake and release capacities are equalized, and either rich or lean from the air-fuel ratio stoichiometry The absorption capacity can be increased against fluctuations on the other side, and the exhaust purification efficiency can be kept at its best.

  Therefore, based on the output of the front air-fuel ratio sensor installed upstream of the catalyst, the oxygen storage amount of the catalyst is obtained by integrating the oxygen excess / deficiency amount (converted from the air-fuel ratio) of the exhaust gas flowing into the catalyst, and this oxygen storage amount The air-fuel ratio is feedback controlled so as to match the target value.

  However, the front air-fuel ratio sensor is subject to deterioration over time due to exposure to high exhaust temperatures, etc. Also, there is a variation in quality at the time of sensor manufacture, and it shifts (shifts) to the stoichiometric point of the air-fuel ratio. In some cases, the stoichiometric point may shift to the rich side or the lean side. If such a stoichiometric shift occurs in the detected air-fuel ratio, the oxygen storage amount of the catalyst calculated based on the output of the front air-fuel ratio sensor becomes inaccurate, and as a result, the oxygen storage amount of the catalyst becomes the target value. It cannot be converged.

  For this reason, Patent Document 1 proposes that a rear air-fuel ratio sensor is also provided downstream of the catalyst, and a shift of the stoichiometric point of the front air-fuel ratio sensor is determined from the output of the rear air-fuel ratio sensor.

This is because the output of the rear air-fuel ratio sensor has been controlled so that the oxygen storage amount becomes the target value based on the output of the front air-fuel ratio sensor. If it is detected, it is determined that the actual oxygen storage amount greatly deviates from the target value, and the cause of this error is that the stoichiometric point of the front air-fuel ratio sensor is shifted to the lean side or the rich side. It certifies that it is. In this case, when the stoichiometric point of the front air-fuel ratio sensor is shifted, it is determined in which direction the stoichiometric point of the front air-fuel ratio sensor is shifted from the air-fuel ratio detected on the downstream side. If it is shifted to the lean side, the stoichiometric point of the front air-fuel ratio sensor is corrected to the lean side accordingly, so that the oxygen storage amount accurately matches the target value.
JP 2003-41990 A

  By the way, in the technique described in Patent Document 1, the stoichiometric point of the rear air-fuel ratio sensor is lean or the control is performed so that the oxygen storage amount becomes the target value based on the output of the front air-fuel ratio sensor. It is determined that the stoichiometric point of the front air-fuel ratio sensor has shifted when a certain time has elapsed since switching to the rich state. The reason that the fixed time has passed is to exclude the case where the front air-fuel ratio sensor's stoichiometric point is temporarily lean or rich due to reasons other than the shift of the stoichiometric point. This is because it is recognized that the stoichiometric point of the front air-fuel ratio sensor is shifted to the lean side or the rich side only when the stoichiometric point of the front air-fuel ratio sensor is shifted to the lean or rich side.

  However, it has been found by subsequent experiments that the phenomenon that the output of the rear air-fuel ratio sensor repeats rich and lean in a short cycle occurs when the stoichiometric point of the front air-fuel ratio sensor is actually lean-shifted.

  This will be further described with reference to FIG. 2. FIG. 2 shows that the oxygen storage amount, the rear air-fuel ratio sensor output, the front air-fuel ratio sensor output when the stoichiometric point of the front air-fuel ratio sensor is lean-shifting when the air amount is low (steady). The model shows how the air-fuel ratio changes. In the uppermost part of FIG. 2, the solid line represents the movement of the calculated oxygen storage amount (oxygen storage amount estimated value), and the broken line represents the movement of the actual oxygen storage amount. Further, in the lowermost stage of FIG. 2, the solid line indicates the movement of the target air-fuel ratio, the alternate long and short dash line indicates the movement of the value obtained by converting the output of the front air-fuel ratio sensor with the stoichiometric point at stoichiometric, and the broken line indicates the stoichiometric point lean shift. The movement of the value obtained by converting the output of the front air-fuel ratio sensor into the air-fuel ratio is shown.

  In FIG. 2, if the calculated oxygen storage amount matches the actual oxygen storage amount, the calculated oxygen storage amount is maintained at the target oxygen storage amount (a value approximately between EMPTY and FULL). Therefore, the rear air-fuel ratio sensor output should not deviate from the stoichiometric lean at the timing t1, but the rear air-fuel ratio sensor output actually deviates from the stoichiometric lean at the timing t1. . This means that the actual oxygen storage amount is not a calculated value, but increases as shown by the broken line and reaches the maximum value (FULL) of the oxygen storage amount at the timing of t1. .

  When the rear air-fuel ratio sensor output deviates from stoichiometric at the timing of t1, the calculated oxygen storage amount is set stepwise to the maximum value (FULL), and the oxygen storage amount set to this maximum value and the target Since the target air-fuel ratio is set based on the deviation from the oxygen storage amount, the target air-fuel ratio goes toward the side where the calculated oxygen storage amount is reduced, that is, the rich side (decreases in steps) at the timing of t1, and this The air / fuel ratio is returned to the rich side by the air / fuel ratio feedback correction coefficient α calculated based on the deviation between the target air / fuel ratio on the rich side and the actual air / fuel ratio (detected by the front air / fuel ratio sensor). To do.

  In this case, the actual air-fuel ratio follows the target air-fuel ratio with a delay, but the actual air-fuel ratio detected by the front air-fuel ratio sensor in which the stoichiometric point is lean-shifted (see the broken line at the bottom of FIG. 2) is Since the stoichiometric point is shifted to the lean side from the actual air-fuel ratio detected by the front air-fuel ratio sensor at the stoichiometric position (refer to the one-dot chain line in the bottom of FIG. 2), this is the lean shift of the stoichiometric point in the first place. The deviation from the target air-fuel ratio increases by the amount of the shift, and therefore the delay of the actual air-fuel ratio with respect to the target air-fuel ratio increases. Therefore, the calculated oxygen storage amount exceeds the target oxygen storage amount and reaches the minimum value ( EMPTY), and at the timing t2, the rear air-fuel ratio sensor output is now deviated to the rich side. At this time, the calculated oxygen storage amount is set stepwise to the minimum value (EMPTY), and the target air-fuel ratio is set based on the deviation between the oxygen storage amount set to the minimum value and the target oxygen storage amount. When the target air-fuel ratio is set, the target air-fuel ratio heads toward the lean side (increases stepwise) at the timing t2, and the air-fuel ratio feedback is calculated based on the deviation between the target air-fuel ratio on the lean side and the actual air-fuel ratio. The correction factor α attempts to return the air / fuel ratio to the lean side this time.

  In this case, since the actual air-fuel ratio follows the target air-fuel ratio with a delay, the calculated oxygen storage amount goes beyond the target oxygen storage amount and reaches the maximum value (FULL), and the rear empty space is reached at the timing t3. The fuel ratio sensor output is now on the lean side.

  After t3, it is the same as after t1, and in this way, the rear air-fuel ratio sensor output repeats lean and rich in a short cycle (cycle from t1 to t3), and the cause is the stoichiometric point of the front air-fuel ratio sensor. Is considered to be in a lean shift.

  Therefore, it is necessary to detect the case where the rear air-fuel ratio sensor output repeats lean and rich in a short cycle and correct the shift of the front air-fuel ratio sensor to the lean side.

  However, when the stoichiometric point of the front air-fuel ratio sensor has shifted to the lean side, the output of the rear air-fuel ratio sensor repeats lean and rich in a short cycle as shown in FIG. According to the technique described in Patent Document 1, the rear air-fuel ratio sensor output reverses to the opposite side before a predetermined time elapses, and the stoichiometric point of the front air-fuel ratio sensor shifts. It may not be detected.

  Therefore, the present invention aims to increase the chances of correcting the stoichiometric point of the front air-fuel ratio sensor so that it can be detected even when the phenomenon shown in FIG. 2 occurs in the output of the rear air-fuel ratio sensor.

  The present invention relates to a front air-fuel ratio in a catalyst having an oxygen storage capability for taking in or releasing oxygen in exhaust gas according to an exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value. The oxygen storage amount is estimated based on the output of the sensor, and the oxygen storage amount is controlled when the rear air-fuel ratio sensor output continues to be lean while the oxygen storage amount estimated value HOSC is controlled to match the target value. The area obtained by integrating the value obtained by subtracting the oxygen storage amount estimated value HOSC from the maximum amount value HOSCMAX is defined as the first area S1, and the oxygen storage amount is estimated from the oxygen storage amount estimated value HOSC when the rear air-fuel ratio sensor output continues to be rich. The area obtained by integrating the value obtained by subtracting the minimum amount HOSCMIN is calculated as the second area S2. Based on the first area S1 and the second area S2, it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted. From this determination result, the stoichiometric point of the front air-fuel ratio sensor is lean-shifted In addition, the stoichiometric point of the front air-fuel ratio sensor is corrected to the lean side.

  Further, the present invention provides a catalyst having an oxygen storage capability for taking in or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value. The oxygen storage amount is estimated based on the output of the fuel ratio sensor, and the air fuel ratio is controlled so that this oxygen storage amount estimated value HOSC coincides with the target value, while the rear air fuel ratio sensor output crosses from the stoichiometric to richer. Sometimes the value obtained by subtracting the oxygen storage amount estimated value HOSC from the oxygen storage amount maximum value HOSCMAX is set as the first length L1, and the oxygen storage amount estimated value when the rear air-fuel ratio sensor output crosses from stoichiometric to lean. The value obtained by subtracting the minimum oxygen storage amount HOSCMIN from HOSC is expressed as the second length L2. Then, based on the first length L1 and the second length L2, it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted. From this determination result, the stoichiometric point of the front air-fuel ratio sensor is lean-shifted. The stoichiometric point of the front air-fuel ratio sensor is corrected to the lean side when the engine is in operation.

  Further, the present invention provides a catalyst having an oxygen storage capability for taking in or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value. When the oxygen storage amount is estimated based on the output of the fuel ratio sensor and the air / fuel ratio is controlled so that the oxygen storage amount estimated value HOSC matches the target value, the rear air / fuel ratio sensor output VRO2 continues to be lean The area obtained by integrating the lean side output is calculated as a third area S3, and the area obtained by integrating the rich side output when the rear air-fuel ratio sensor output VRO2 continues to be rich is calculated as a fourth area S4. And the fourth area S4, it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted. The stoichiometric point of the front air-fuel ratio sensor is configured to correct the lean side when the stoichiometric point of the cement the air-fuel ratio sensor is lean shift.

Further, the present invention provides a catalyst having an oxygen storage capability for taking in or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value. The oxygen storage amount is estimated based on the output of the fuel ratio sensor, and the air / fuel ratio is controlled so that the estimated value of the oxygen storage amount coincides with the target value, while the rear air / fuel ratio sensor output VRO2 is the minimum value per certain section. Amplitude count CNT1 that swings to a maximum value or from a maximum value to a minimum value is calculated, and it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the amplitude count CNT1 per fixed section.
From this determination result, when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted, the stoichiometric point of the front air-fuel ratio sensor is corrected to the lean side.

  Further, the present invention provides a catalyst having an oxygen storage capability for taking in or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value. The oxygen storage amount is estimated based on the output of the fuel ratio sensor, and the air fuel ratio is controlled so that the oxygen storage amount estimated value HOSC matches the target value, while the stoichiometric point of the front air fuel ratio sensor is lean-shifted. A rich side slice level GRSLICE that sometimes does not cross the rear air-fuel ratio sensor output VRO2, and a lean-side slice level GLSLICE that the rear air-fuel ratio sensor output VRO2 crosses when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted, The rear air-fuel ratio sensor output VRO2 becomes the rich side slice per fixed interval. The number of times CNT2 that continuously crosses the lean side slice level GLSLICE in the same direction without crossing the bell GRSLICE is calculated, and the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the number of continuous crossings CNT2 per fixed section. From this determination result, when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted, the stoichiometric point of the front air-fuel ratio sensor is corrected to the lean side.

  According to the present invention, the front air-fuel ratio sensor also has a phenomenon in which the output of the rear air-fuel ratio sensor repeats rich and lean in a short cycle due to the lean shift of the stoichiometric point of the front air-fuel ratio sensor. It is possible to detect that the stoichiometric point is lean-shifted, thereby increasing the opportunity to correct the stoichiometric point of the front air-fuel ratio sensor.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows a schematic configuration of an exhaust emission control device to which the present invention is applied. A catalyst 3 is provided in an exhaust pipe 2 of an engine 1, a front air-fuel ratio sensor 4 having a linear characteristic upstream, and a rear air-fuel ratio sensor 4 downstream. An air-fuel ratio sensor 5 is installed, and a controller 6 that controls the air-fuel ratio of the fuel supplied to the engine 1 based on the output of these sensors is provided.

  The intake pipe 7 of the engine 1 is provided with a throttle valve 8 and an air flow meter 9 for measuring the intake air amount adjusted by the throttle valve 8.

  The catalyst 3 is a so-called three-way catalyst, and purifies NOx, HC, and CO with maximum efficiency when the catalyst atmosphere has a stoichiometric air-fuel ratio. The catalyst 3 is coated on the catalyst carrier with an oxygen storage material such as ceria, and has a function (oxygen storage function) of holding or releasing oxygen according to the air-fuel ratio of the inflowing exhaust gas.

  The front air-fuel ratio sensor 4 has a linear output characteristic corresponding to the oxygen concentration in the exhaust gas, and the rear air-fuel ratio sensor 5 has a characteristic of switching in a binary manner near the stoichiometric range.

  Further, the engine 1 is provided with a temperature sensor 10 for detecting the temperature of the cooling water, and is used for determining the activated state of the catalyst 3 as well as the operating state of the engine 1.

  The controller 6 comprises a microprocessor, ROM, RAM, I / O interface, etc., and calculates the oxygen storage amount of the catalyst 3 based on the output of the air flow meter 9 and the front air-fuel ratio sensor 4, and this storage amount is the target value. The air-fuel ratio is feedback controlled so that

  The calculation of the oxygen storage amount of the catalyst 3 is estimated from the exhaust air-fuel ratio upstream of the catalyst and the intake air amount at that time. First, an excess oxygen ratio, which is a ratio of excess or deficiency of oxygen in the exhaust gas, is calculated from the exhaust air / fuel ratio upstream of the catalyst. The oxygen excess rate is zero when stoichiometric, and is positive on the lean side and negative on the rich side.

  The amount of oxygen absorbed or released by the catalyst 3 is known from the oxygen excess rate and the amount of intake air at that time, and the amount of oxygen stored in the catalyst 3 can be estimated by integrating this. The maximum oxygen storage amount of the catalyst 3 is confirmed in advance by experiments or the like, for example, a half of the retention amount is set as a target value, and the air-fuel ratio is controlled so that the oxygen storage amount matches this target value.

  When the calculated oxygen storage amount is smaller than the target value, the target air-fuel ratio is made leaner to increase the holding amount, and conversely when it is larger than the target value, the oxygen storage amount is decreased to reduce the oxygen storage amount. To match.

  However, the target value of the air-fuel ratio is also controlled from the engine operation request, and is basically feedback controlled so as to be stoichiometric, so even if the deviation from the target storage amount is corrected, the correction Regarding the amount, the correction range for the stoichiometry is limited to a range that does not deteriorate the drivability.

  Further, although there is a deviation between the calculated oxygen storage amount and the actual oxygen storage amount due to the calculation error, based on the oxygen concentration detected by the rear air-fuel ratio sensor 5, for example, when the engine fuel is cut, After shifting to fuel cut, the oxygen storage amount calculated at a predetermined timing is reset to correct the deviation.

  Furthermore, in the present invention, the controller 6 determines whether or not the stoichiometric point of the front air-fuel ratio sensor 4 for calculating the oxygen storage amount is shifted to the lean side, and if the sensor air is deteriorated, the stoichiometric point of the front air-fuel ratio sensor 4 is determined. When the engine is lean-shifting, the stoichiometric point of the front air-fuel ratio sensor 4 is corrected to the lean side accordingly, and the oxygen storage amount is correctly converged to the target value.

  Since the oxygen storage amount of the catalyst 3 is controlled so as to always become the target value, even if there is some fluctuation in the air-fuel ratio on the upstream side of the catalyst, the air-fuel ratio on the downstream side of the catalyst becomes stoichiometric due to the oxygen storage capacity of the catalyst. Maintained. However, if the stoichiometric point of the front air-fuel ratio sensor 4 is shifted to the lean side, the oxygen storage amount of the catalyst is saturated, and the rear air-fuel ratio sensor output becomes the lean side from the stoichiometric. Thereafter, due to the delay of the actual air-fuel ratio with respect to the target air-fuel ratio, a phenomenon occurs in which the rear air-fuel ratio sensor output repeats rich and lean in a short cycle (see FIG. 2).

  It is determined whether or not such a phenomenon has occurred, and when it is determined that such a phenomenon has occurred, it is determined that the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted, and the front air-fuel ratio sensor 4 The stoichiometric point is corrected to the lean side.

  This determination will be described with reference to FIG. 3. When the lean shift occurs in the left half of FIG. 3, the change in the rear air-fuel ratio sensor output and the oxygen storage amount is similarly changed to the right half of FIG. This shows the change in the rear air-fuel ratio sensor output and the oxygen storage amount when no oxygen occurs. Note that the left half of FIG. 3 is basically the same as FIG. 2, except that the time scale of the horizontal axis is different.

  As shown in FIG. 3, when the rear air-fuel ratio sensor output continues to be lean, the area (indicated by S1) surrounded by the oxygen storage amount maximum value HOSCMAX and the oxygen storage amount current value HOSC (new) is “first”. “Area”, an area (indicated by S2) surrounded by the oxygen storage amount current value HOSC (new) and the oxygen storage amount minimum value HOSCMIN (= 0) when the rear air-fuel ratio sensor output continues to be rich. Considering these area ratios RATESMO2 (= first area / second area) as “area”, if the stoichiometric point of the front air-fuel ratio sensor is lean-shifted, this area ratio RATESMO2 exceeds 1 (FIG. Note that the area ratio RATESMO2 is 1 or less (see the right half of FIG. 3) when there is no lean shift. Determines the stoichiometric point of the front air-fuel ratio sensor 4 whether or not the lean shift.

  This control content will be described in detail according to the following flowchart.

  FIG. 4 is for calculating (estimating) the oxygen storage amount HCSC, and is executed at regular intervals (for example, every 10 ms).

  First, in Step 1, it is checked whether or not the catalyst 3 is activated according to conditions such as cooling water temperature. If the catalyst 3 is not activated, the oxygen storage capability of the catalyst 3 does not work, so the current process is terminated.

  If the catalyst 3 is activated, the process proceeds to step 2 where the output of the front air-fuel ratio sensor 4 is converted into the air-fuel ratio AFs. For example, a table showing the relationship between the output of the front air-fuel ratio sensor 4 and the air-fuel ratio may be created in advance, and the air-fuel ratio may be obtained by searching this table.

  In step 4, the shift correction amount learned value is read from a memory such as an EEPROM, and in step 5, shift correction to the lean side of the stoichiometric point of the front air-fuel ratio sensor 4 is performed based on the shift correction amount learned value. This is to calculate the actual air-fuel ratio AFr by the following equation.

AFr = AFs + shift correction amount learning value (1)
Here, AFs on the right side of the equation (1) is an apparent air-fuel ratio detected by the front air-fuel ratio sensor 4, and when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, the actual air-fuel ratio is the apparent air-fuel ratio. Since it is on the lean side, the difference is made up by the shift correction amount learning value. Therefore, when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, the shift correction amount learning value is a positive value, and the side on which the air-fuel ratio becomes larger by learning this positive value of the shift correction amount (that is, the lean side). It is corrected to. The calculation of the shift correction amount learning value will be described later.

  In step 6, the excess / deficiency oxygen concentration FO2 in the exhaust gas is obtained by searching a predetermined table from the actual air-fuel ratio AFr which is the air-fuel ratio after shift correction thus obtained.

  Here, as shown in FIG. 5, the excess / deficient oxygen concentration FO2 in the exhaust gas is a value obtained by converting the air-fuel ratio at that time into an oxygen concentration with the stoichiometric value as a reference zero. Therefore, for example, when the air-fuel ratio is lean, it exceeds the stoichiometric oxygen concentration, so FO2 is a positive value. When the air-fuel ratio is rich, it is insufficient than the stoichiometric oxygen concentration. It becomes.

  In step 7, the oxygen storage amount HOSC is calculated based on this excess / deficiency oxygen concentration FO2. The calculation of the oxygen storage amount HOSC will be described with reference to the flowcharts of FIGS. 6 and 7 (subroutine of step 7 in FIG. 4).

  The oxygen storage characteristic of the catalyst 3 is divided into a characteristic of being absorbed / released at a high speed by the noble metal of the catalyst 3 and a characteristic of being absorbed / released at a low speed by an oxygen storage material such as ceria of the catalyst 3. Therefore, the actual storage amount according to the characteristics of the catalyst 3 can be accurately calculated by calculating the oxygen storage amount separately for the high speed component and the low speed component in accordance with this characteristic.

  FIG. 6 is for calculating the high speed component of the oxygen storage amount, and FIG. 7 is for calculating the low speed component.

  In FIG. 6, in step 11, the oxygen excess / deficiency O2IN per calculation cycle time is calculated by the following equation.

O2IN = a * FO2 * Q * t (2)
Where a: constant (value for unit conversion),
Q: Exhaust flow rate (substitute with intake air flow rate),
t: calculation cycle time of FIG.
In step 12, it is determined whether the high-speed component oxygen is absorbed or the high-speed component oxygen is released based on the oxygen excess / deficiency amount O2IN. If the air-fuel ratio of the exhaust gas flowing into the catalyst 3 is lean and the oxygen excess / deficiency O2IN is greater than zero, it is determined that the high-speed component oxygen is absorbed, and the routine proceeds to step 13 where Is calculated.

HO2 (new) = HO2 (old) + O2IN (3)
However, HO2 (new): current value of high-speed component,
HO2 (old): the previous value of the high speed component,
On the other hand, when it is determined that the oxygen excess / deficiency amount O2IN is equal to or less than zero and the high-speed component oxygen is being released, the process proceeds from step 12 to step 14 to calculate the high-speed component by the following equation.

HO2 (new) = HO2 (old) + O2IN × A (4)
However, HO2 (new): current value of high-speed component,
HO2 (old): the previous value of the high speed component,
A: Oxygen release rate of high-speed component,
If the high-speed component current value HO2 (new) is calculated in this way, whether the value does not exceed the maximum value HO2MAX of the high-speed component in steps 15 and 16, or is not less than the minimum value HO2MIN (= 0). See. If the high speed component current value HO2 (new) is greater than or equal to the maximum value HO2MAX, the process proceeds from step 15 to step 17 and the overflow amount (excess amount) OVERFLOW overflowing without being absorbed as a high speed component is calculated by the following equation. Furthermore, the high speed component current value HO2 (new) is limited to the maximum value HO2MAX.

OVERFLOW = HO2 (new) -HO2MAX (5)
If the high speed component current value HO2 (new) is less than or equal to the minimum value HO2MIN, the process proceeds from step 16 to step 18, and the overflow amount (insufficient amount) OVERFLOW overflowing without being absorbed as the high speed component is calculated by the following equation. Furthermore, the high speed component current value HO2 (new) is limited to the minimum value HO2MIN.

OVERFLOW = HO2 (new) -HO2MIN (6)
Here, since 0 is given as the minimum value HO2MIN, the amount of oxygen deficient in a state where all of the high-speed component oxygen is released is calculated as a negative overflow amount.

  On the other hand, when the high speed component current value HO2 (new) is between the maximum value HO2MAX and the minimum value HO2MIN, the oxygen excess / deficiency O2IN of the exhaust gas flowing into the catalyst 3 is all absorbed as high speed component oxygen. Proceeding to step 19, the overflow OVERFLOW is set to zero.

  In steps 20 and 22, it is checked whether the rear air-fuel ratio sensor output has changed from stoichiometric to lean, or whether the rear air-fuel ratio sensor output has changed from stoichiometric to rich. When the rear air-fuel ratio sensor output changes from stoichiometric to lean, the routine proceeds from step 20 to step 21 where the high-speed component current value HO2 (new) changes to the maximum value HO2MAX, and the rear air-fuel ratio sensor output changes from stoichiometric to richer. Sometimes the routine proceeds from step 20 or 22 to step 23, where the high speed component current value HO2 (new) is set to the minimum value HO2MIN. Thus, when the rear air-fuel ratio sensor output changes from stoichiometric to lean, the high-speed component current value HO2 (new) is set to the maximum value HO2MAX, and when the rear air-fuel ratio sensor output changes from stoichiometric to richer The reason why the high speed component current value HO2 (new) is set to the minimum value HO2MIN is to match the high speed component current value HO2 (new), which is the calculated oxygen storage amount, with the actual oxygen storage amount.

  The overflow component OVERFLOW overflowing from the high-speed component current value HO2 (new) above the maximum value HO2MAX or below the minimum value HO2MIN without being high-speed oxygen is absorbed or released as low-speed oxygen. FIG. 7 is a subroutine for calculating the low-speed component of the oxygen storage amount. In this subroutine, the low-speed component is calculated based on the overflow amount OVERFLOW that overflows without becoming the high-speed oxygen component.

  In FIG. 7, at step 41, the low speed component is calculated by the following equation.

LO2 (new) = LO2 (old) z + OVERFLOW × B (7)
However, LO2 (new): current value of low speed component,
LO2 (old): previous value of low speed component,
B: Oxygen absorption / release rate of low-speed component,
Here, the oxygen absorption / release rate B of the low-speed component is set to a positive value of 1 or less, but actually has different characteristics between absorption and release, and the actual absorption / release rate is the catalyst temperature TCAT, the low-speed component. Since it is affected by components and the like, the absorption rate and the release rate may be set separately. In that case, oxygen is excessive when the overflow amount OVERFLOW is positive, and as the oxygen absorption rate B at this time, for example, a larger value is set as the catalyst temperature TCAT is higher and the low-speed component LO2 is smaller. Further, when the overflow amount OVERFLOW is negative, oxygen is insufficient. As the oxygen release rate B at this time, for example, a larger value is set as the catalyst temperature TCAT is higher and the low-speed component is larger.

  In steps 42 and 43, the calculated low-speed component current value LO2 (new) does not exceed the maximum value LO2MAX, or does not fall below the minimum value LO2MIN (= 0), as in the high-speed component calculation. See how. If the maximum value LO2MAX is exceeded, the routine proceeds from step 42 to step 44, where the oxygen excess / deficiency O2OUT overflowing from the low speed component current value LO2 (new) is calculated according to the following equation, and the low speed component current value LO2 (new) is calculated as the maximum value LO2MAX. Restrict to.

O2OUT = LO2 (new) -LO2MAX (8)
The oxygen excess / deficiency amount O2OUT is oxygen flowing out downstream of the catalyst 3 as it is.

  On the other hand, if it is below the minimum value LO2MIN, the process proceeds from step 43 to step 45, where the low speed component current value LO2 (new) is limited to the minimum value LO2MIN.

  In this way, the oxygen storage amount of the catalyst 3 is calculated.

  FIG. 8 is for calculating the air-fuel ratio feedback correction coefficient α, and is executed each time a Ref signal (crank angle reference position signal) is input to each cylinder. This is in accordance with the fact that fuel injection is performed in synchronization with the Ref signal, and the air-fuel ratio in the exhaust gas changes due to this fuel injection.

  In step 51, it is determined whether or not the air-fuel ratio feedback control condition is satisfied. When the air-fuel ratio feedback control condition is satisfied, it is the same as the conventional case. For example, when the front air-fuel ratio sensor 4 is activated, the air-fuel ratio feedback control condition is satisfied. Also, the air-fuel ratio feedback control condition is not satisfied when the fuel is cut or the engine is heavily loaded. When the air-fuel ratio feedback control condition is not satisfied, the routine proceeds to step 52, where the air-fuel ratio feedback correction coefficient α is set to 1.0 and the current processing is terminated.

  When the air-fuel ratio feedback control condition is satisfied, the routine proceeds from step 51 to step 53, where the high-speed component current value HO2 (new) obtained in the flow of FIG. 6 and the actual air-fuel ratio obtained in step 5 of FIG. AFr is read, and in step 54, the high speed component current value HO2 (new) is transferred to the oxygen storage amount HOSC. 6 and 7, the oxygen storage amount HOSC has been explained by calculating separately for the high speed component and the low speed component. However, for the sake of simplicity, the high speed component current value HO2 (new) is used as the oxygen storage amount HOSC.

  In step 55, the deviation HOSCS of the oxygen storage amount HOSC from the target oxygen storage amount is calculated by the following equation.

HOSCS = HOSC−target oxygen storage amount (9)
Here, ½ of the maximum oxygen storage amount is set as the target oxygen storage amount of equation (9).

  In step 56, based on this deviation HOSCS, the target oxygen storage amount FO2t is calculated by the following equation.

FO2t = Gp × HOSCS (10)
Where Gp: proportional gain (applicable value),
Here, the proportional gain Gp in the equation (10) is introduced in order to cope with a case where there is a demand to increase the response to the target air-fuel ratio. Therefore, Gp = 1.0 may be set when there is no need to increase the response to the target air-fuel ratio.

  In step 57, the target oxygen storage amount FO2t is converted to the target air-fuel ratio AFt by the following equation.

AFt = 14.7 + 14.7 × FO2t / Q (11)
However, Q: exhaust flow rate (substitute with intake air flow rate),
Here, the second term on the right side of the equation (11) is a value obtained by converting the target oxygen storage amount FO2t into an air-fuel ratio, and the equation (11) is a value obtained by shifting the theoretical air-fuel ratio (14.7) by this air-fuel ratio conversion. Is set as the target air-fuel ratio. For example, when the oxygen storage amount HOSC is larger than the target value (deviation HOCSS is positive), the target oxygen storage amount FO2t becomes a positive value, and the target air-fuel ratio AFt becomes larger than 14.7 according to the equation (11) (stoichiometric). Than the lean side). That is, when the oxygen storage amount HOSC is larger than the target value, the target air-fuel ratio is shifted to a leaner side than the stoichiometric state in order to return to the target value. Conversely, when the oxygen storage amount HOSC is smaller than the target value (deviation HOSCS is negative), the target oxygen storage amount FO2t becomes a negative value, and the target air-fuel ratio AFt is smaller than 14.7 according to the equation (11). (Rich side than stoichiometric). That is, when the oxygen storage amount HOSC is smaller than the target value, the target air-fuel ratio is shifted to a richer side than the stoichiometric value in order to return to the target value.

  Since the air-fuel ratio control here is feedback control, in step 58, the deviation dAF of the actual air-fuel ratio AFr detected by the front air-fuel ratio sensor 4 from the target air-fuel ratio AFt is calculated by the following equation.

dAF = AFr−AFt (12)
However, the actual air-fuel ratio AFr detected by the front air-fuel ratio sensor 4 on the right side of the equation (12) is a value corrected to the lean side by the shift correction amount learning value as shown in the equation (1).

  In step 59, an air-fuel ratio feedback correction coefficient α is calculated by the following equation based on the air-fuel ratio deviation dAF.

α = dAF × PGain + ΣdAF × IGain + ΔdAF × DGain
... (13)
Where PGain: proportional gain (positive value)
IGain: integral gain (positive value),
DGain: differential gain (positive value),
The first term, the second term, and the third term on the right side of the equation (13) are the proportional component, the integral component, and the derivative component, respectively, and the equation (13) is the sum of these as the feedback correction amount of the air-fuel ratio.

  In a fuel injection amount calculation routine (not shown), the fuel injection pulse width Ti at the time of sequential injection is calculated by the following equation using the air-fuel ratio feedback correction coefficient α calculated in this way.

Ti = Tp × TFBYA × α × 2 + Ts (14)
Where Tp: basic injection pulse width,
TFBYA: target equivalent ratio,
Ts: Invalid injection pulse width,
When the predetermined injection timing is reached, the fuel injectors provided in the respective cylinders are opened by this fuel injection pulse width Ti, and fuel injection is performed.

  When the air-fuel ratio feedback control condition is satisfied (in this case, TFBYA = 1.0), for example, when the actual air-fuel ratio AFr is larger than the target air-fuel ratio AFt (the air-fuel ratio deviation dAF is positive), the air-fuel ratio feedback correction coefficient α is 1.0. The value exceeding this 1.0 is increased by the value of α exceeding 1.0, and the amount of fuel obtained from the stoichiometric mixture is increased, and the air-fuel ratio in the exhaust is returned to stoichiometric. Conversely, when the actual air-fuel ratio AFr is smaller than the target air-fuel ratio AFt (the air-fuel ratio deviation dAF is negative), the air-fuel ratio feedback correction coefficient α is a positive value smaller than 1.0, and a positive value smaller than 1.0 As a result, the air fuel ratio in the exhaust gas is returned to stoichiometric by reducing the fuel amount from which the stoichiometric mixture is obtained. That is, when the air-fuel ratio feedback control condition is satisfied, fuel control is performed so that the actual air-fuel ratio AFr detected by the front air-fuel ratio sensor 4 matches the target air-fuel ratio AFt. In this case, if the feedback gain (PGain, IGain, DGain) on the right side of the equation (13) is excessively increased, the actual air-fuel ratio AFr exceeds the target air-fuel ratio AFt, so-called overshoot occurs, so the feedback gain is too large. Therefore, when the air-fuel ratio feedback control is performed, the actual air-fuel ratio AFr follows the target air-fuel ratio AFt with a response delay.

  FIGS. 9 and 10 are used to calculate a shift correction amount learning value used to correct the stoichiometric point of the front air-fuel ratio sensor 4 to the lean side, and is executed at regular intervals. Here, a case where the shift correction amount is a learning value will be described, but it is needless to say that the shift correction amount may not be a learning value.

  In step 61, it is checked whether or not the operating condition determined from the engine speed and the basic injection pulse width Tp is in the learning region. The region for calculating (updating) the shift correction amount learning value is predetermined as shown in FIG. The reason why the high air amount region is not used as the learning region is that the phenomenon shown in the left half of FIG. 2 or FIG. 3 hardly occurs in the high air amount region. If the operating condition is not in the learning area, the current process is terminated.

  When the driving condition is in the learning region, the process proceeds from step 61 to step 62 and subsequent steps. Among these, steps 62 to 70 in FIG. 9 and steps 71 to 75 and 80 in FIG. 10 are portions for calculating the first area S1 and the second area S2 shown in FIG. 3, steps 74 to 77 and 79 to in FIG. 85 is a part for calculating the first area average value AEVS1 and the second area average value AVES2, step 86 in FIG. 10 is a part for calculating the area ratio RATESMO2, and step 87 in FIG. 10 is the stoichiometric point of the front air-fuel ratio sensor 4. The step 88 of FIG. 10 for determining whether or not the lean shift is performed is based on the determination result, and the shift correction amount learning value is set to a positive value when the stoichiometric point of the front air-fuel ratio sensor 4 is lean shifted. This is the part that is updated on the larger side.

  Here, the average value of the first area and the average value of the second area are calculated, and the ratio of these is calculated as an area ratio. The ratio may be calculated as the area ratio.

  This will be specifically described below. In FIG. 9, first, in step 62, it is determined whether or not the rear air-fuel ratio sensor output is stoichiometric.

  Here, the rich, lean and stoichiometric determinations based on the output of the rear air-fuel ratio sensor having hysteresis are as follows. That is, as shown in FIG. 23, when the rear air-fuel ratio sensor output goes from the rich side to the lean side, the rear air-fuel ratio sensor output is on the rich side before crossing the rich side slice level RSLICE. It is determined that it has returned to stoichiometry at a timing when it has become smaller across the level RSLICE. On the contrary, when the rear air-fuel ratio sensor output goes from the lean side to the rich side, if the rear air-fuel ratio sensor output crosses the lean side slice level LSLICE, it is on the lean side, and increases across the lean side slice level LSLICE. It is determined that it has returned to stoichiometric timing.

  When the rear air-fuel ratio sensor output is not stoichiometric (ie, rich or lean), the routine proceeds to step 63, where the oxygen storage amount HOSC is read. As the oxygen storage amount HOSC, the high speed component HO2 current value (new) may be used as described in steps 53 and 54 of FIG.

  In steps 64 and 65, it is checked whether or not the rear air-fuel ratio sensor output is lean this time, and whether or not the rear air-fuel ratio sensor output was lean last time. When the rear air-fuel ratio sensor output is lean this time and was also lean last time, that is, when the rear air-fuel ratio sensor output continues to be lean, the routine proceeds from step 64 to step 66 to step 66, and the following equation shows the In order to calculate one area S1, the value obtained by subtracting the oxygen storage amount HOSC from the maximum value HOSCMAX of the oxygen storage amount is integrated.

LSMOSQ (new) = LSMOSQ (old) + (HOSCMAX-HOSC)
... (15)
However, LSMOSQ (new): Lean area this time,
LSMOSQ (old): previous area value when lean,
In step 67, the value of the current area value LSMOSQ (new) at the time of leaning is moved to the previous value LSMOSQ (old) at the time of leaning for the next processing.

  After the next time, if the rear air-fuel ratio sensor output is lean, the operations in steps 66 and 67 are repeated. As a result, according to equation (15), while the rear air-fuel ratio sensor output is continuously lean, the second term on the right side of equation (15) is integrated, so the area S1 shown in FIG. 3 is obtained. Will be. The initial value of the previous area value LSMOSQ (old) at the time of lean is set to zero.

  On the other hand, when the rear air-fuel ratio sensor output is rich this time, the routine proceeds from step 64 to step 68 to check whether the rear air-fuel ratio sensor output was rich last time. When the rear air-fuel ratio sensor output is rich this time and was also rich last time, that is, when the rear air-fuel ratio sensor output continues to be rich, the routine proceeds from step 64, 68 to step 69, and the following equation shows the In order to calculate the two areas S2, a value obtained by subtracting the minimum value HOSCMIN (= 0) of the oxygen storage amount from the oxygen storage amount HOSC is integrated.

RSMOSQ (new) = RSMOSQ (old) + HOSC-HOSCMIN
... (16)
However, RSMOSQ (new): Rich area area value,
RSMOSQ (old): previous area value when rich,
In step 70, the value of the rich area current value RSMOSQ (new) is transferred to the rich area previous value RSMOSQ (old) for the next processing.

  If the rear air-fuel ratio sensor output is rich after the next time, the operations in steps 69 and 70 are repeated. As a result, according to the equation (16), while the rear air-fuel ratio sensor output is continuously rich, the second term on the right side of the equation (16) is integrated, so the area during rich is shown in FIG. The second area S2 is obtained. Zero is also set to the initial value of the previous area value RSMOSQ (old) when rich.

  On the other hand, when the rear air-fuel ratio sensor output is currently stoichiometric, the routine proceeds from step 62 in FIG. 9 to steps 71 and 72 in FIG. 10 to check whether the rear air-fuel ratio sensor output was the last stoichiometric or the previous lean. . When the rear air-fuel ratio sensor output is currently stoichiometric and last lean, that is, when the rear air-fuel ratio sensor output is switched from lean to stoichiometric, the routine proceeds from step 71 to step 73 to the current lean area at this time The value LSMOSQ (new) is moved to the area S1, and in step 74, the lean area is integrated by the following equation.

LSUM1 (new) = LSUM1 (old) + S1 (17)
However, LSUM1 (new): Lean area integrated value this value,
LSUM1 (old): Lean area integrated value previous value,
Here, the area S1 of the second term on the right side of the equation (17) is a value at the timing when the rear air-fuel ratio sensor output is switched from lean to stoichiometric, and is therefore the first area S1 itself shown in FIG. Therefore, equation (17) is for further integrating the first area S1. This is for obtaining the average value after sampling and integrating several first areas S1.

  In step 75, the sample count m1 (initially set to zero) is increased by one. The number of times of sampling m1 represents the number of times that the first area S1 is integrated according to the equation (17). In step 76, the value of the lean area integrated value current value LSUM1 (new) is moved to the lean area integrated value previous value LSUM1 (old) for the next processing. The initial value of the lean area integrated value previous value LSUM1 (old) is set to zero.

  In step 77, in order to obtain the first area S1 shown in FIG. 3 again in preparation for when the rear air-fuel ratio sensor output continues to be lean next time, the previous value LSMOSQ (old) at the time of leaning is set to zero.

  On the other hand, when the rear air-fuel ratio sensor output is currently stoichiometric and rich last time, that is, when the rear air-fuel ratio sensor output is switched from rich to stoichiometric, the routine proceeds from step 71 and 72 to step 78. The current value RSMOSQ (new) at the time of rich is moved to the area S2, and the area at the time of rich is integrated by the following equation at step 79.

RSUM1 (new) = RSUM1 (old) + S2 (18)
However, RSUM1 (new): Rich area integrated value this value,
RSUM1 (old): Rich area integrated value previous value,
Here, the area S2 in the second term on the right side of the equation (18) is the value at the timing when the rear air-fuel ratio sensor output is switched from rich to stoichiometric, and is therefore the second area S2 itself shown in FIG. Therefore, the equation (18) further accumulates the second area S2. This is to obtain an average value after sampling and integrating several second areas S2.

  In step 80, the number of samples n1 (initially set to zero) is increased by one. The number of times of sampling n1 represents the number of times that the second area S2 is integrated according to the equation (18). In step 81, the value of the rich area integrated value current value RSUM1 (new) is moved to the rich area integrated value previous value RSUM1 (old) for the next processing. The initial value of the rich area integrated value previous value RSUM1 (old) is also set to zero.

  In step 82, in order to obtain the second area S2 shown in FIG. 3 again in preparation for the next time when the rear air-fuel ratio sensor output continues to be rich, the rich area previous value RSMOSQ (old) = 0 is set.

  In this way, each time the rear air-fuel ratio sensor output is switched from lean to stoichiometric, the first area S1 shown in FIG. 3 is integrated in step 74, and the rear air-fuel ratio sensor output is switched from rich to stoichiometric. In each step 79, the second area S2 shown in FIG. 3 is accumulated.

  In step 83, the sample counts m1 and n1 are compared with a constant value (adapted value). If the number of samples m and n has not reached a certain value, the current process is terminated.

  If the number of samples m1 and n1 both reach a certain value, the process proceeds to steps 84 and 85, and the average value AVES1 of the first area S1 and the average value AVES2 of the second area S2 are calculated by the following equations, respectively.

AVES1 = LSUM1 (new) / m1 (19)
AVES2 = RSUM1 (new) / n1 (20)
In step 86, these area ratios RATESMO2 are calculated by the following equation.

RATESMO2 = AVES1 / AVES2 (21)
In step 87, the area ratio RATESMO2 is compared with 1 (a constant value). When the area ratio RATESMO2 exceeds 1, it is determined that the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted, and the routine proceeds to step 88 where the shift correction amount learning value (initially set to zero) is increased according to the following equation. Update.

Shift correction amount learning value (new) = Shift correction amount learning value (old) + predetermined value
... (22)
However, shift correction amount learning value (new): shift correction amount learning value before update,
Shift correction amount learned value (old): Updated shift correction amount learned value,
Predetermined value: Positive constant value (adapted value) that determines the amount of update
The shift correction amount learning value thus obtained is stored in a memory such as an EEPROM.

  On the other hand, if the area ratio RATESMO2 is 1 or less, the stoichiometric point of the front air-fuel ratio sensor 4 has not been lean shifted, so at this time step 88 is skipped and the routine proceeds to step 89.

  In step 89, in order to prepare for the next update of the shift correction amount learning value, the lean area integrated value previous value LSUM1 (old), the rich area integrated value previous value RSUM1 (old), and the sampling times m1 and n1 are all zero. Put it in.

  Here, the operation of the present embodiment will be described.

  When the stoichiometric point of the front air-fuel ratio sensor 4 is not lean-shifted, the shift correction amount learning value = 0, so that the value AFs obtained by converting the output of the front air-fuel ratio sensor 4 to the air-fuel ratio is in good agreement with the actual air-fuel ratio. Will do. However, due to deterioration over time, as shown in the left half of FIG. 3, a phenomenon occurs in which the output of the rear air-fuel ratio sensor 5 repeats lean and rich in a short cycle and the center of amplitude is biased to the lean side. This is because the front air-fuel ratio sensor 4 apparently outputs an output on the rich side at the stoichiometric point. Therefore, the shift correction amount learning value may be corrected to the lean side by adding to the apparent air-fuel ratio AFs. .

  Therefore, when the phenomenon shown in the left half of FIG. 3 occurs, it is determined that the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted, and the shift correction amount learning value is updated to a positive value and increased. The shift correction amount learning value has a positive value. At this time, the value AFs obtained by converting the front air-fuel ratio sensor output to the air-fuel ratio is corrected by the shift correction amount learning value in step 5 of FIG. (The shift correction amount learning value is added to the value AFs obtained by converting the front air-fuel ratio sensor output into the air-fuel ratio), and the actual air-fuel ratio AFr increases. This correction causes the actual air-fuel ratio AFr as the calculated value to approach the actual air-fuel ratio. However, if the phenomenon shown in the left half of FIG. 3 still occurs, the correction to the lean side is insufficient. At this time, the shift correction amount learning value is updated again to a larger positive value, the shift correction amount learning value becomes larger than the previous value, and the AFs is corrected by the increased shift correction amount learning value. Then, the actual air-fuel ratio AFr as the calculated value further approaches the actual air-fuel ratio. In this way, when the update of the shift correction amount learning value is repeated, the actual air-fuel ratio AFr as the calculated value eventually converges to the actual air-fuel ratio. At this time, as shown in the right half of FIG. 3, the phenomenon that the rear air-fuel ratio sensor output repeats rich and lean in a short cycle is stopped. In this way, the shift correction amount learned value is updated many times every predetermined time as long as the operating condition stays in the learning region, so that the shift correction amount learned value converges.

  Thus, according to the present embodiment (the invention described in claims 1 and 2), the oxygen storage amount is estimated based on the output of the front air-fuel ratio sensor 4, and the oxygen storage amount estimated value HOSC is the target value. A value obtained by subtracting the oxygen storage amount estimated value HOSC from the oxygen storage amount maximum value HOSCMAX when the rear air-fuel ratio sensor output continues to be lean, while controlling the air-fuel ratio to coincide with Is the first area S1, and when the rear air-fuel ratio sensor output continues to be rich, the area obtained by subtracting the oxygen storage amount minimum value HOSCMIN from the oxygen storage amount estimated value HOSC is the second area. Each is calculated as S2 (see steps 73 and 78 in FIG. 10), and the front air-fuel ratio is calculated based on the first area S1 and the second area S2. It is determined whether or not the stoichiometric point of the sensor 4 is lean-shifted (see steps 84, 85, 86 and 87 in FIG. 10), and when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on this determination result Since the stoichiometric point of the front air-fuel ratio sensor 4 is corrected to the lean side (see steps 87 and 88 in FIG. 10 and steps 4 and 5 in FIG. 4), the stoichiometric point of the front air-fuel ratio sensor 4 is caused by the lean shift. Even when the phenomenon that the output of the rear air-fuel ratio sensor 5 repeats lean and rich in a short cycle occurs, it is possible to detect that the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted. Opportunities for correcting the stoichiometric point of the front air-fuel ratio sensor 4 can be increased.

  Further, according to the first embodiment (the invention described in claim 2), since the partner to be compared with the area ratio RATESMO2 may be one, it is not necessary to bother the partner to be compared with the area ratio RATESMO2, and the configuration is simple. It has become.

  The flowcharts of FIGS. 13 and 14 are the second embodiment, which replaces FIGS. 9 and 10 of the first embodiment. 13 and 14, the same step numbers are assigned to the same portions as those in FIGS. 9 and 10.

  In the first embodiment, as shown in the lower half of the left half of FIG. 3, the average values AVES1 and AVES2 are obtained after sampling and integrating the first area S1 and the second area S2 with respect to the oxygen storage amount. The area ratio RATESMO2 is calculated from the two average values AVES1 and AVES2, and it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted depending on whether or not the area ratio RATESMO2 exceeds 1. In the embodiment, as shown in the left half of FIG. 12 (basically the same as FIG. 3), the maximum oxygen storage amount HOSCMAX when the rear air-fuel ratio sensor output is switched from stoichiometric to rich and the oxygen at that time The difference (indicated by L1) from the storage amount HOSC is “first length”, and the rear air-fuel ratio sensor output is switched from stoichiometric to lean. The difference between the oxygen storage amount HOSC and the minimum value HOSCMIN (= 0) (indicated by L2) is “second length”, and these ratios RATEO2 (= first length / second length) are When considered, when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, this ratio RATEO2 exceeds 1 (see the left half of FIG. 12), whereas when the lean-shift does not occur, the ratio Focusing on the fact that RATEO2 is 1 or less (see the right half of FIG. 12), it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted depending on whether the ratio RATEO2 exceeds 1. is there.

  That is, steps 62, 71, 63, 64, 111, and 116 in FIG. 13 are portions for calculating the first length L1 and the second length L2 shown in FIG. 12, and steps 112 to 115 and 117 to 120 in FIG. And steps 121 to 123 in FIG. 14 calculate the first length average value AEVL1 and second length average value AVEL2, step 124 in FIG. 14 calculates the ratio RATEO2, and step 125 in FIG. A portion for determining whether or not the stoichiometric point of the fuel ratio sensor 4 is in a lean shift, step 88 in FIG. 14 is based on the determination result, and the shift correction amount when the stoichiometric point of the front air-fuel ratio sensor 4 is in a lean shift. This is the part where the learning value is updated to a larger value with a positive value.

  However, here, the average value of the first length and the average value of the second length are calculated, and the case where these ratios are calculated will be described, but the first length and the second length are simply calculated, These ratios may be calculated.

  When the difference from the first embodiment is mainly described, in steps 62, 71 and 64 in FIG. 13, it is determined whether the rear air-fuel ratio sensor output is at the current stoichiometric, the previous stoichiometric or at the current lean. When the rear air-fuel ratio sensor output is not stoichiometric at this time, is the last stoichiometric and is not lean at this time, that is, when the rear air-fuel ratio sensor output is switched from stoichiometric to rich, the routine proceeds from step 62, 71, 64 to step 111. The value obtained by subtracting the current oxygen storage amount value HOSC from the maximum oxygen storage amount value HOSCMAX is shifted to the length L1, and the length at the time of rich switching is added by the following equation at step 112.

RSUM2 (new) = RSUM2 (old) + L1 (23)
However, RSUM2 (new): Rich switching length integrated value this value,
RSUM2 (old): Rich change length integrated value previous value,
Here, the length L1 of the second term on the right side of the equation (23) is a value at the timing when the rear air-fuel ratio sensor output is switched from the stoichiometric state to the rich state, and therefore the first length L1 itself shown in FIG. is there. Therefore, the equation (23) further accumulates the first length L1. This is for obtaining an average value after sampling and integrating several first lengths L1.

  In step 113, the number of samples m2 (initially set to zero) is increased by one. The number of times of sampling m2 represents the number of times that the first length L1 is integrated according to the equation (23). In step 114, the value of the rich switching length integrated value current value RSUM2 (new) is moved to the rich switching length integrated value previous value RSUM2 (old) for the next processing. Zero is set in the initial value of the rich integrated time length integrated value previous value RSUM2 (old).

  On the other hand, when the rear air-fuel ratio sensor output is not the current stoichiometric, the previous stoichiometric and the current lean, that is, when the rear air-fuel ratio sensor output is switched from stoichiometric to lean, the steps 62, 71, 64 Proceeding to step 116, the value obtained by subtracting the minimum value HOSCMIN (= 0) of the oxygen storage amount from the oxygen storage amount HOSC at that time is moved to the length L2, and in step 117, the length at the time of lean switching is integrated by the following equation.

LSUM2 (new) = LSUM2 (old) + L2 (24)
However, LSUM2 (new): Accumulated value at the time of lean switching This value,
LSUM2 (old): Lean switching length integrated value previous value,
Here, since the length L2 of the second term on the right side of the equation (24) is a value at the timing when the rear air-fuel ratio sensor output is switched from stoichiometric to lean, it is the second length L2 itself shown in FIG. is there. Therefore, the equation (24) further accumulates the second length L2. This is for obtaining an average value after sampling and integrating several second lengths L2.

  In step 118, the number of samples n2 (initially set to zero) is increased by one. The number of samples n2 represents the number of times that the second length L2 has been integrated according to the equation (24). In step 119, the value of the current value LSUM2 (new) at the time of lean switching is moved to the previous value LSUM2 (old) at the time of the lean switching for the next processing. The initial value of the length integrated value previous value LSUM2 (old) at the time of lean switching is also set to zero.

  In this way, every time the rear air-fuel ratio sensor output is switched from stoichiometric to rich, the first length L1 shown in FIG. 12 is integrated in step 112 of FIG. 13, and the rear air-fuel ratio sensor output is leaner than stoichiometric. Each time the switch is made, the second length L2 shown in FIG. 12 is accumulated in step 117 of FIG.

  In FIG. 14, in step 121, the number of samples m2 and n2 is compared with a constant value (adapted value). If the number of samples m2 and n2 has not reached a certain value, the current process is terminated.

  If both the number of samples m2 and n2 reach a constant value, the process proceeds to steps 122 and 123, and the average value AVEL1 of the first length L1 and the average value AVEL2 of the second length L2 are calculated by the following equations.

AVEL1 = LSUM2 (new) / m2 (25)
AVEL2 = RSUM2 (new) / n2 (26)
In step 124, the ratio RATEO2 is calculated by the following equation.

RATEO2 = AVEL1 / AVEL2 (27)
In step 125, the ratio RATEO2 is compared with 1. When the ratio RATEO2 is greater than 1, it is determined that the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, and the routine proceeds to step 88 where the shift correction amount learning value is increased by a positive value according to the above equation (22). Update to The shift correction amount learning value thus obtained is stored in a memory such as an EEPROM.

  On the other hand, if the area ratio RATESMO2 is 1 or less, the stoichiometric point of the front air-fuel ratio sensor 4 has not been lean shifted, so at this time step 88 is skipped and the routine proceeds to step 126.

  In step 126, in order to prepare for the next update of the shift correction amount learning value, the rich switching length integrated value previous value LSUM2 (old), the lean switching length integrated value previous value RSUM2 (old), the number of samples m2, and n2 are set. Both are set to zero.

  As described above, according to the second embodiment (the inventions according to claims 3 and 4), the oxygen storage amount is estimated based on the output of the front air-fuel ratio sensor, and this oxygen storage amount estimated value becomes the HOSC target value. While controlling the air-fuel ratio so as to match, the first length is a value obtained by subtracting the oxygen storage amount estimated value HOSC from the oxygen storage amount maximum value HOSCMAX when the rear air-fuel ratio sensor output crosses from stoichiometric to richer Further, when the rear air-fuel ratio sensor output crosses from stoichiometric to lean, the values obtained by subtracting the oxygen storage amount minimum value HOSCMIN from the oxygen storage amount estimated value HOSC are respectively calculated as the second lengths (step 111 in FIG. 13). 116), the stoichiometric point of the front air-fuel ratio sensor 4 is lean shifted based on the first length and the second length. (See steps 122, 123, 124, and 125 in FIG. 14), and when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the determination result, the stoichiometric point of the front air-fuel ratio sensor 4 is determined. Is corrected to the lean side (see Steps 125 and 88 in FIG. 14 and Steps 4 and 5 in FIG. 4), as in the first embodiment, due to the lean shift of the stoichiometric point of the front air-fuel ratio sensor 4, It is possible to detect that the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted even when the rear air-fuel ratio sensor outputs a phenomenon in which lean and rich are repeated in a short cycle. Opportunities for correcting the stoichiometric point of the fuel ratio sensor 4 can be increased.

  In this case, a method for determining whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted is simpler than in the first embodiment.

  Also in the second embodiment (the invention described in claim 4), similarly to the first embodiment, the number of opponents to be compared with the ratio RATEO2 may be 1, so that it is necessary to bother the counterparts to be compared with the ratio RATEO2. It has a simple configuration.

  The flowcharts of FIGS. 16 and 17 are the third embodiment, the flowchart of FIG. 19 is the fourth embodiment, and the flowchart of FIG. 21 is the fifth embodiment, which replaces FIGS. 9 and 10 of the first embodiment, respectively. . In FIG. 16, FIG. 17, FIG. 19, and FIG. 21, the same steps as those in FIG. 9 and FIG.

  In the first and second embodiments, as shown in each left half of FIGS. 3 and 12, it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the oxygen storage amount. However, in the third, fourth, and fifth embodiments, the front air-fuel ratio sensor 4 is based on the rear air-fuel ratio sensor output itself as shown in the left halves of FIGS. It is determined whether or not the stoichiometric point is lean-shifted. That is, in the third embodiment, as shown in the left half of FIG. 15, the area (indicated by S3) when the rear air-fuel ratio sensor output is continuously lean is “third area”, and the rear air-fuel ratio sensor output is When the area when continuously rich (indicated by S4) is “fourth area” and these area ratios RATERO2 (= third area / fourth area) are considered, the stoichiometric point of the front air-fuel ratio sensor 4 This area ratio RATERO2 exceeds a certain value when lean shift occurs (see the left half of FIG. 15), whereas the area ratio RATERO2 does not exceed a certain value when no lean shift occurs (see FIG. 15). Paying attention to the right half of FIG. 15), after sampling and integrating several third areas S3 and fourth areas S4, the average values AVES3 and AVES4 are obtained, and the two average values AVE are obtained. 3, the area ratio RATERO2 calculated from AVES4, determines whether the whether the area ratio RATERO2 exceeds a certain value stoichiometric point of the front air-fuel ratio sensor 4 is lean shift.

  In this case, the constant value that is the counterpart to be compared with the area ratio RATERO2 is a value larger than 1 unlike the first and second embodiments.

  In the fourth embodiment, as shown in the left half of FIG. 18, when the operating condition stays in the learning region continuously, the rear air-fuel ratio sensor output from the minimum value to the maximum value and the minimum value from the maximum value per certain section. The amplitude number CNT1 swung up to is calculated, and it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on whether or not the amplitude number CNT1 exceeds the reference value 1 (adapted value).

  In the fifth embodiment, as shown in the left half of FIG. 20, apart from the normal slice levels (RSLICE, LSLICE), the learning slice level (learning rich side slice level GRSLICE and learning lean side slice level GLSLICE). And the learning lean side slice level GLSLICE does not cross the learning rich side slice level GRSLICE among the learning slice levels in the same direction (in the figure, It is determined whether or not the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted based on whether or not the vehicle has continuously crossed a certain number of times. Here, when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, the learning rich-side slice level GRSLICE is at a position where the rear air-fuel ratio sensor output VRO2 does not cross, and the learning lean-side slice level GLSLICE. Two learning slice levels (GRSLICE and GLSLICE) are set in advance at a position where the rear air-fuel ratio sensor output VRO2 crosses.

  Explaining from the third embodiment, steps 62, 131, 64, 65, 68, 132 to 135 in FIG. 16 and steps 71, 72, 135, and 140 in FIG. 17 correspond to the third area S3 and the fourth area shown in FIG. A part for calculating the area S4, a part for calculating the third area average value AEVS3 and the fourth area average value AVES4 for steps 136 to 139 and 141 to 147 in FIG. 17, and a part for calculating the area ratio RATERO2 in step 148 in FIG. Step 149 in FIG. 17 is a part for determining whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted. Step 88 in FIG. 17 is based on the determination result, and the stoichiometric point of the front air-fuel ratio sensor 4 is lean. In this case, the shift correction amount learning value is updated to a larger value with a positive value when shifting.

  However, here, the average value of the third area and the average value of the fourth area are calculated, and the case where the ratio of these areas is calculated will be described. The area ratio may be calculated.

  In the third embodiment, the differences from the first embodiment will be mainly described. When the rear air-fuel ratio sensor output is not stoichiometric in FIG. 16, the routine proceeds from step 62 to step 131, and the rear air-fuel ratio sensor output VRO2 is read. This rear air-fuel ratio sensor is a sensor that changes in a binary manner as in the first embodiment.

  In steps 64 and 65, it is checked whether or not the rear air-fuel ratio sensor output is lean this time, and whether or not the rear air-fuel ratio sensor output was lean last time. When the rear air-fuel ratio sensor output is lean this time and was also lean last time, that is, when the rear air-fuel ratio sensor output continues to be lean, the routine proceeds from step 64, 65 to step 132, and the following equation shows the In order to calculate the three areas S3, the value obtained by subtracting the rear air-fuel ratio sensor output from the lean side slice level LSLICE is integrated.

LSMRO2 (new) = LSMRO2 (old) + (LSLICE-VRO2)
... (28)
However, LSMRO2 (new): Lean area this time,
LSMRO2 (old): Lean area previous value,
LSLICE: Lean side slice level (normal),
In step 133, the value of the current area LSMRO2 (new) at the time of leaning is moved to the previous value LSMRO2 (old) at the time of leaning for the next processing.

  After the next time, if the rear air-fuel ratio sensor output is lean, the operations of steps 132 and 133 are repeated. As a result, according to the equation (28), the second term on the right side of the equation (28) is integrated while the rear air-fuel ratio sensor output is continuously lean, so the area S3 shown in FIG. 15 is obtained. Will be. The initial value of the previous area value LSMRO2 (old) at the time of lean is set to zero.

  On the other hand, when the rear air-fuel ratio sensor output is rich this time, the routine proceeds from step 64 to step 68 to check whether the rear air-fuel ratio sensor output was rich last time. When the rear air-fuel ratio sensor output is rich this time and has been rich last time, that is, when the rear air-fuel ratio sensor output continues to be rich, the routine proceeds from step 64 to step 134 to step 134, and the following equation shows the In order to calculate the four areas S4, a value obtained by subtracting the rich slice level RSLICE from the rear air-fuel ratio sensor output VRO2 is integrated.

RSMRO2 (new) = RSMRO2 (old) + (VRO2-RSLICE)
... (29)
However, RSMRO2 (new): Rich area area value,
RSMRO2 (old): Rich previous area value,
In step 135, the value of the current value RSMRO2 (new) at the time of rich is moved to the previous value RSMRO2 (old) at the time of rich for the next processing.

  If the rear air-fuel ratio sensor output is rich after the next time, the operations of steps 134 and 135 are repeated. As a result, according to the equation (29), while the rear air-fuel ratio sensor output is continuously rich, the second term on the right side of the equation (29) is integrated, so the fourth area S4 shown in FIG. Will be required. Zero is also set to the initial value of the previous area value RSMRO2 (old) when rich.

  On the other hand, when the rear air-fuel ratio sensor output is currently stoichiometric, the routine proceeds from step 62 in FIG. 16 to steps 71 and 72 in FIG. 17 to see whether the rear air-fuel ratio sensor output was last stoichiometric or whether it was last lean. . When the rear air-fuel ratio sensor output is at the current stoichiometric time and was lean last time, that is, when the rear air-fuel ratio sensor output is switched from lean to stoichiometric, the routine proceeds from step 71 to step 135 to the current lean area at this time The value LSMRO2 (new) is moved to the area S3, and in step 136, the lean area is integrated by the following equation.

LSUM3 (new) = LSUM3 (old) + S3 (30)
However, LSUM3 (new): Lean area integrated value this value,
LSUM3 (old): Lean area integrated value previous value,
Here, the area S3 in the second term on the right side of the equation (30) is the value at the timing when the rear air-fuel ratio sensor output is switched from lean to stoichiometric, and is therefore the third area S3 itself shown in FIG. Therefore, the equation (30) further accumulates the third area S3. This is for obtaining an average value after sampling and integrating several third areas S3.

  In step 137, the number of samples m3 (initially set to zero) is increased by one. The number of samples m3 represents the number of times that the third area S3 has been integrated according to the equation (30). In step 138, the value of the lean area integrated value current value LSUM3 (new) is moved to the lean area integrated value previous value LSUM3 (old) for the next processing. The initial value of the lean area integrated value previous value LSUM3 (old) is set to zero.

  In step 139, in order to obtain the third area S3 shown in FIG. 15 again in preparation for the next time that the rear air-fuel ratio sensor output continues to be lean, the previous area value LSMRO2 (old) at lean time is set to zero.

  On the other hand, when the rear air-fuel ratio sensor output is currently stoichiometric and rich last time, that is, when the rear air-fuel ratio sensor output is switched from rich to stoichiometric, the routine proceeds from step 71 and 72 to step 140. The current value RSMRO2 (new) of the rich time is moved to the area S4, and the rich time area is integrated by the following equation in step 141.

RSUM3 (new) = RSUM3 (old) + S4 (31)
However, RSUM3 (new): Rich area integrated value this value,
RSUM3 (old): Rich area integrated value previous value,
Here, the area S4 in the second term on the right side of the equation (31) is a value at the timing when the rear air-fuel ratio sensor output is switched from rich to stoichiometric, and is the fourth area S4 itself shown in FIG. Therefore, the equation (31) further accumulates the fourth area S4. This is to obtain an average value after sampling and integrating several fourth areas S4.

  In step 142, the number of samples n3 (initially set to zero) is increased by one. The number of samples n3 represents the number of times that the fourth area S4 has been integrated according to the equation (31). In step 143, the value of the rich area integrated value current value RSUM3 (new) is moved to the rich area integrated value previous value RSUM3 (old) for the next processing. The initial value of the rich area integrated value previous value RSUM3 (old) is also set to zero.

  In step 144, in order to obtain the fourth area S4 shown in FIG. 15 again in preparation for the next time when the rear air-fuel ratio sensor output continues to be rich, the rich area previous value RSMRO2 (old) = 0 is set.

  In this way, every time the rear air-fuel ratio sensor output is switched from lean to stoichiometric, the third area S3 shown in FIG. 15 is integrated in step 136, and the rear air-fuel ratio sensor output is switched from rich to stoichiometric. In step 141, the fourth area S4 shown in FIG.

  In step 145, the sample counts m3 and n3 are compared with a constant value (adapted value). If the number of samples m3 and n3 has not reached a certain value, the current process is terminated.

  If both the number of samples m3 and n3 reach a constant value, the process proceeds to steps 146 and 147, and the average value AVES3 of the third area S3 and the average value AVES4 of the fourth area S4 are calculated by the following equations.

AVES3 = LSUM3 (new) / m3 (32)
AVES4 = RSUM3 (new) / n3 (33)
In step 148, these area ratios RATERO2 are calculated by the following equation.

RATERO2 = AVES3 / AVES4 (34)
In step 149, the area ratio RATERO2 is compared with a constant value (adapted value). Here, the reason why the partner to be compared with the area ratio RATERO2 in the third embodiment is set to a constant value (a value exceeding 1) instead of 1 is due to the difference in the waveform of the rear air-fuel ratio sensor output and the oxygen storage amount ( (See the difference between the top and bottom of the left half of Figure 3). When the area ratio RATERO2 is larger than a certain value, it is determined that the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, and the routine proceeds to step 88 where the shift correction amount learning value is increased by a positive value by the above equation (22). Update to the side. The shift correction amount learning value thus obtained is stored in a memory such as an EEPROM.

  On the other hand, if the area ratio RATERO2 is equal to or less than a certain value, the stoichiometric point of the front air-fuel ratio sensor 4 has not been lean shifted, so at this time step 88 is skipped and the routine proceeds to step 150.

  In step 150, in order to prepare for the next update of the shift correction amount learning value, all zeros are entered in the lean area integrated value previous value LSUM3 (old), the rich area integrated value previous value RSUM3 (old), and the sampling times m3 and n3. Keep it.

  The shift correction amount learning value is updated many times every predetermined time as long as the driving condition stays in the learning region. As a result, the shift correction amount learned value eventually converges, and the value (AFs) obtained by converting the output of the front air-fuel ratio sensor 4 to the air-fuel ratio by the converged learned value is corrected to be increased by the above equation (1). Even if the stoichiometric point of the front air-fuel ratio sensor 4 is actually in a lean shift state, the corrected actual air-fuel ratio AFr is a front air-fuel ratio sensor in which the stoichiometric point is not lean-shifted. It becomes the same as the air-fuel ratio detected by 4. Therefore, after the shift correction amount learning value has converged, the phenomenon shown in the left half of FIG. 15, that is, the phenomenon that the rear air-fuel ratio sensor output repeats lean and rich in a short cycle and the center of the amplitude is biased to the lean side. As shown in the right half of FIG.

  Thus, according to the third embodiment (the invention described in claim 5), the oxygen storage amount is estimated based on the output of the front air-fuel ratio sensor 4, and the oxygen storage amount estimated value HOSC coincides with the target value. Thus, while controlling the air-fuel ratio, the area when the rear air-fuel ratio sensor output VRO2 continues to be lean is set as the third area S3, and the area when the rear air-fuel ratio sensor output VRO2 continues to be rich is set to the fourth area S3. Each area S4 is calculated (see steps 135 and 140 in FIG. 17), and it is determined whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the third area and the fourth area (FIG. 17). Steps 146, 147, 148, and 149), the front air-fuel ratio is detected when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the determination result. Since the stoichiometric point of the sensor 4 is corrected (see Steps 149 and 88 in FIG. 17 and Steps 4 and 5 in FIG. 4), it is caused by the lean shift of the stoichiometric point of the front air-fuel ratio sensor 4 as in the first embodiment. Thus, it is possible to detect that the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted even when the output of the rear air-fuel ratio sensor repeats lean and rich in a short cycle. Opportunities for correcting the stoichiometric point of the front air-fuel ratio sensor 4 can be increased.

  Next, the fourth embodiment will be described. In FIG. 19, steps 163 to 164 are portions for determining whether or not the driving conditions are kept in the learning region, and steps 165 to 168 are the driving conditions being continued. Step 169 is a part for calculating the number of amplitudes CNT1 in which the rear air-fuel ratio sensor output swings from the minimum value to the maximum value or from the maximum value to the minimum value per predetermined section when staying in the learning region, step 169. Step 88 determines whether or not the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted. Step 88 is based on the determination result and shifts when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted. This is a part for updating the correction amount learning value to a larger value with a positive value.

  Specifically, in FIG. 19, when the driving condition is in the learning region, the process proceeds from step 61 to step 161 to set the flag F1 = 1, and when the driving condition is not in the learning area, the process proceeds from step 61 to step 162 and the flag F1 = 0. To do. A flag F1 = 1 indicates that the driving condition is in the learning region, and a flag F1 = 0 indicates that the driving condition is not in the learning region.

  In steps 163 and 164, the value of the flag F1 and the value of the flag F1 (previous) are checked. Here, the flag F1 (previous) contains the previous value of the flag F1. When the flag F1 = 1 and F1 (previous) = 0, that is, when the flag F1 is switched from 0 to 1 (when the driving condition enters the learning area from outside the learning area), from steps 163 and 164 In Steps 165 and 166, the number of amplitudes CNT1 = 0 is set, and a timer is started (t = 0). The timer is for measuring the time of staying in the learning area continuously.

  In step 172, the value of the flag F1 is moved to the flag F1 (previous) for the next processing.

  When the flag F1 = 1 and F1 (previous) = 1, that is, when the flag F1 continues to be 1 (when the driving condition continues to stay in the learning region), the process proceeds from step 163, 164 to step 167, and the amplitude The number of times CNT1 is calculated. Although the operation here is not shown, the rear air-fuel ratio sensor output is monitored as shown in the left half of FIG. 18, and the amplitude number CNT1 is increased by 1 when the rear air-fuel ratio sensor output fluctuates from the minimum value to the maximum value. In addition, when the rear air-fuel ratio sensor output fluctuates from the maximum value to the minimum value, the amplitude count CNT1 is increased by 1, and these operations are repeatedly executed. The left half of FIG. 18 shows the number of times of amplitude five times.

  In step 168, the timer value t is compared with a certain time. The fixed time is a value (adapted value) for determining a fixed section in FIG. If the timer value t is less than or equal to a certain time, steps 169 to 171 are skipped, the operation of step 172 is executed, and the current process is terminated.

  If the operating condition continues to stay in the learning region, the operations of Steps 163, 164, 167, and 168 will be repeated from the next time, and the stoichiometric point of the front air-fuel ratio sensor 4 may be lean-shifted by this operation. In this case, the number of amplitudes CNT1 increases. Eventually, when the timer value t exceeds a certain time, the routine proceeds from step 168 to step 169, where the number of amplitudes CNT1 is compared with the reference value 1 (adapted value). If the number of amplitudes CNT1 exceeds the reference value 1, it is determined that the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted, and the routine proceeds to step 88 where the shift correction amount learning value is set to a positive value according to the above equation (22). Update to the larger value. The shift correction amount learning value thus obtained is stored in a memory such as an EEPROM. This completes the first update of the shift correction amount learning value.

  On the other hand, if the number of amplitudes CNT1 is less than or equal to the reference value 1, the stoichiometric point of the front air-fuel ratio sensor 4 has not been lean shifted. At this time, step 88 is skipped and the routine proceeds to step 170.

  In steps 170 and 171, in order to prepare for the second shift correction amount learning value update, the number of amplitudes CNT1 = 0 is set and the timer is restarted (t = 0), and then the operation of step 172 is executed. As a result, as long as the driving condition continues to stay in the learning region, the shift correction amount learning value is updated each time a certain time elapses.

  On the other hand, if the operating condition deviates from the learning region before the timer value t exceeds a certain time, the process proceeds from step 61, 162, 163 to steps 173, 174, and the number of amplitudes CNT1 = 0 is set and the timer is reset ( After t = 0), the operation of step 172 is executed.

  Thus, according to the fourth embodiment (the inventions described in claims 7 and 8), the oxygen storage amount is estimated based on the output of the front air-fuel ratio sensor 4, and this oxygen storage amount estimated value HOSC is the target value. While controlling the air-fuel ratio so as to coincide with each other, the number of amplitudes CNT1 at which the rear air-fuel ratio sensor output swings from the minimum value to the maximum value or from the maximum value to the minimum value per certain section is calculated (steps 167 and 168 in FIG. 19). (Refer to steps 168 and 169 in FIG. 19) based on the number of amplitudes CNT1 per fixed section, and determine whether or not the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted. When the stoichiometric point of the fuel ratio sensor 4 is lean-shifted, the stoichiometric point of the front air-fuel ratio sensor 4 is corrected to the lean side (step in FIG. 19). 69, 88, steps 4 and 5 in FIG. 4), the output of the rear air-fuel ratio sensor leans in a short cycle due to the lean shift of the stoichiometric point of the front air-fuel ratio sensor 4 as in the first embodiment. It is possible to detect that the stoichiometric point of the front air-fuel ratio sensor 4 has been lean-shifted even when the phenomenon of repeating the rich and the rich has occurred, thereby providing an opportunity to correct the stoichiometric point of the front air-fuel ratio sensor 4 Can be increased.

  Next, the fifth embodiment will be described. FIG. 21 of the fifth embodiment is obtained by replacing the number of amplitudes CNT1 with the number of continuous crossings CNT2 in FIG. 19 of the fourth embodiment (see steps 181 to 185 in FIG. 21). ).

  In this case, in the fifth embodiment, as shown in FIG. 20, a learning slice level (GRSLICE and GLSLICE) is provided separately from the normal slice level (RSLICE and LSLICE). In FIG. 21, steps 182 and 168 indicate that the front air-fuel ratio sensor does not cross the newly provided learning rich side slice level GRSLICE and is newly provided another learning lean side slice level per predetermined interval. The step of calculating the number of continuous crossings CNT2 that continuously crosses GLSLICE from stoichiometric to lean, step 183 determines whether the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the number of continuous crossings CNT2. Step 88 is a portion for updating the shift correction amount learning value to a side that increases by a positive value when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the determination result.

  However, here, a case where the front air-fuel ratio sensor calculates the number of continuous crossings CNT2 that continuously crosses the lean side slice level GLSLICE for learning from stoichiometric to lean will be described. Even when GLSLICE is continuously calculated from the lean to the stoichiometric number of continuous crossings, it may be calculated.

  Specifically, in the fifth embodiment as well, in the same way as in the fourth embodiment, when the operation condition continues to stay in the learning region, the process proceeds from step 163 and 164 to step 182 to calculate the number of consecutive crossings CNT2. The calculation of the number of continuous crossings CNT2 will be described with reference to the flowchart of FIG.

  In FIG. 22 (subroutine of step 182 of FIG. 21), the rear air-fuel ratio sensor output VRO2 is read in step 191 and the rear air-fuel ratio sensor output VRO2 and the learning rich side slice level GRSLICE are compared in step 192. When the rear air-fuel ratio sensor output VRO2 is equal to or higher than the learning rich side slice level GRSLICE (the rear air-fuel ratio sensor output VRO2 is rich), the routine proceeds to step 200, where the number of continuous crossings CNT2 = 0 is set and the current processing is terminated.

  When the rear air-fuel ratio sensor output VRO2 is less than the learning rich side slice level GRSLICE, it is determined that the rear air-fuel ratio sensor output VRO2 is stoichiometric or lean and the routine proceeds to step 193, where the rear air-fuel ratio sensor output VRO2 and the learning lean side slice The level GLSLICE is compared with step 192. When the rear air-fuel ratio sensor output VRO2 exceeds the lean side slice level GLSLICE for learning, it is determined that the rear air-fuel ratio sensor output VRO2 is stoichiometric and the routine proceeds to step 194 where the flag F2 = 1 is set, and the rear air-fuel ratio sensor output VRO2 is When the lean side slice level for learning GLSLICE is equal to or lower, it is determined that the rear air-fuel ratio sensor output VRO2 is lean, and the routine proceeds from step 193 to step 195 to set the flag F2 = 0. A flag F2 = 1 indicates that the rear air-fuel ratio sensor output VRO2 is stoichiometric, and a flag F2 = 0 indicates that the rear air-fuel ratio sensor output VRO2 is lean.

  In steps 196 and 197, the value of this flag F2 and the value of another flag F2 (previous) are checked. The flag F2 (previous) contains the previous value of the flag F2. When the flag F2 = 0 and F2 (previous) = 1, that is, when the flag F2 is switched from 1 to 0, it is determined that the rear air-fuel ratio sensor output VRO2 has crossed from stoichiometric to lean, and from step 196, 197 to step 198 Go ahead and increase the number of continuous crossings CNT2 (initially set to zero) by one.

  In step 199, the value of the flag F2 is moved to the flag F2 (previous) for the next processing.

  In FIG. 21, when the operating condition stays continuously in the learning region, if the front air-fuel ratio sensor 4 has a lean shift of the stoichiometric point by calculating the number of continuous crossings CNT2 in step 198, the number of continuous crossings CNT2 Will increase continuously. If the rear air-fuel ratio sensor output VRO2 exceeds the learning rich side slice level GRSLICE for some reason even if the number of continuous crossings CNT2 has increased so far, the routine proceeds to step 200 in FIG. Therefore, only when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, the number of consecutive crossings CNT2 increases continuously.

  Eventually, when the timer value t exceeds a certain time, the process proceeds from step 168 to step 183 in FIG. 21, and the number of consecutive crossings CNT2 is compared with the reference value 2 (adapted value). If the number of continuous crossings CNT2 exceeds the reference value 2, it is determined that a lean shift has occurred at the stoichiometric point of the front air-fuel ratio sensor 4, and the routine proceeds to step 88 where the shift correction amount learning value is calculated by the above equation (22). Update to a larger value with a positive value. The shift correction amount learning value thus obtained is stored in a memory such as an EEPROM. This completes the update of the first learning value.

  On the other hand, if the number of continuous crossings CNT2 is less than or equal to the reference value 2, the stoichiometric point of the front air-fuel ratio sensor 4 has not been lean shifted, so at this time step 88 is skipped and the routine proceeds to step 184.

  In steps 184 and 171, in order to prepare for the update of the second shift correction amount learning value, the number of consecutive crossings CNT2 = 0 is set and the timer is restarted (t = 0), and then the operation of step 172 is executed. As a result, as long as the driving condition continues to stay in the learning region, the shift correction amount learning value is updated each time a certain time elapses.

  On the other hand, if the operating condition deviates from the learning area before the timer value t exceeds a certain time, the process proceeds from step 61, 162, 163 to steps 185, 174, and the number of continuous crossings is set to CNT2 = 0 and the timer is reset. After (t = 0), the operation of step 172 is executed.

  Thus, according to the fifth embodiment (inventions according to claims 9 and 10), the rear air-fuel ratio sensor output VRO2 does not cross the learning rich side slice level GRSLICE per fixed section and the learning lean side slice. The number of times CNT2 that continuously traverses the level GLSLICE in the same direction is calculated (see steps 191 to 199 in FIG. 22), and the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted based on the number of times of continuous crossing CNT2 per fixed section. (See steps 168 and 183 in FIG. 21), and from this determination result, when the stoichiometric point of the front air-fuel ratio sensor 4 is lean-shifted, the stoichiometric point of the front air-fuel ratio sensor is set to the lean side. (Refer to Steps 183 and 88 in FIG. 21 and Steps 4 and 5 in FIG. 4). Similarly, the front air-fuel ratio sensor 4 also has a phenomenon in which the lean output of the rear air-fuel ratio sensor repeats lean and rich in a short cycle due to the lean shift of the stoichiometric point of the front air-fuel ratio sensor 4. It is possible to detect that the stoichiometric point is lean-shifted, thereby increasing the chance of correcting the stoichiometric point of the front air-fuel ratio sensor 4.

  In the first to third embodiments, it is not necessarily a condition that the driving condition stays continuously in the learning area only if the driving condition is in the learning area. As in the embodiment, it is possible to make it a condition that the driving condition stays in the learning area continuously.

  The function of the air-fuel ratio control means according to claim 1 is according to the flow of FIGS. 4 and 8, the function of the first area calculation means is according to step 73 of FIG. 10, and the function of the second area calculation means is step 78 of FIG. Thus, the function of the judging means is performed by steps 84, 85, 86, 87 in FIG. 10, and the function of the lean side correcting means is performed by steps 87, 88 in FIG. 10 and steps 4, 5 in FIG.

  The function of the air-fuel ratio control means according to claim 3 is the flow of FIGS. 4 and 8, the function of the first length calculation means is the step 111 of FIG. 13, and the function of the second length calculation means is the function of FIG. In step 116, the function of the determination means is performed in steps 122, 123, 124, and 125 in FIG. 14, and the function of the lean side correction means is performed in steps 125 and 88 in FIG. 14 and steps 4 and 5 in FIG. Yes.

  The function of the air-fuel ratio control means according to claim 5 is the flow of FIGS. 4 and 8, the function of the third area calculation means is according to step 135 of FIG. 17, and the function of the fourth area calculation means is step 140 of FIG. Therefore, the function of the determination means is performed by steps 146, 147, 148, and 149 in FIG. 17, and the function of the lean side correction means is performed by steps 149 and 88 in FIG. 17 and steps 4 and 5 in FIG.

  The function of the air-fuel ratio control means according to claim 7 is according to the flow of FIG. 4 and FIG. 8, the function of the amplitude number calculating means is according to steps 167 and 168 of FIG. The function of the lean side correcting means is performed by steps 169 and 88 in FIG. 19 and steps 4 and 5 in FIG.

  The function of the air-fuel ratio control means according to claim 9 is based on the flow of FIGS. 4 and 8, the function of the continuous crossing number calculation means is based on steps 182 and 168 of FIG. 21, and the flow of FIG. 21, the function of the lean side correcting means is performed by steps 183 and 88 in FIG. 21 and steps 4 and 5 in FIG.

1 is a schematic configuration diagram of an embodiment of an engine to which the present invention is applied. Waveform that shows how the oxygen storage amount, rear air-fuel ratio sensor output, and front air-fuel ratio change when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted when the air volume is low (steady) Figure. The wave form diagram showing the change at the time of lean shift about the rear air-fuel-ratio sensor output and oxygen storage amount at the time of no lean shift. The flowchart for demonstrating the calculation of oxygen storage amount. The characteristic view which shows the relationship between the air fuel ratio detected by a front air fuel ratio sensor, and excess-deficient oxygen concentration. The flowchart for demonstrating the calculation of the high-speed component of oxygen storage amount. The flowchart for demonstrating the calculation of the low speed component of oxygen storage amount. The flowchart for demonstrating the calculation of an air fuel ratio feedback correction coefficient. The flowchart for demonstrating the calculation of a shift correction amount learning value. The flowchart for demonstrating the calculation of a shift correction amount learning value. Learning area diagram. The wave form diagram showing the change at the time of the lean shift about the rear air-fuel-ratio sensor output of 2nd Embodiment, and the amount of oxygen storage when there is no lean shift. The flowchart for demonstrating the calculation of the shift correction amount learning value of 2nd Embodiment. The flowchart for demonstrating the calculation of the shift correction amount learning value of 2nd Embodiment. The wave form diagram showing the change at the time of the lean shift about the rear air-fuel-ratio sensor output of 3rd Embodiment at the time of the lean shift absence. The flowchart for demonstrating the calculation of the shift correction amount learning value of 3rd Embodiment. The flowchart for demonstrating the calculation of the shift correction amount learning value of 3rd Embodiment. The wave form diagram showing the change at the time of the lean shift about the rear air-fuel-ratio sensor output of 4th Embodiment at the time of no lean shift. The flowchart for demonstrating the calculation of the shift correction amount learning value of 4th Embodiment. The wave form diagram showing the change at the time of lean shift about the rear air fuel ratio sensor output of 5th Embodiment at the time of no lean shift. The flowchart for demonstrating the calculation of the shift correction amount learning value of 5th Embodiment. The flowchart for demonstrating the calculation of the number of times of continuous crossing. The characteristic diagram of a rear air fuel ratio sensor output.

Explanation of symbols

1 Engine 3 Catalyst 4 Front air-fuel ratio sensor 5 Rear oxygen sensor 6 Controller

Claims (16)

In a catalyst capable of storing or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value,
A front air-fuel ratio sensor installed upstream of the catalyst;
A rear air-fuel ratio sensor installed downstream of the catalyst;
Air-fuel ratio control means for estimating an oxygen storage amount based on the output of the front air-fuel ratio sensor and controlling the air-fuel ratio so that the oxygen storage amount estimated value matches a target value;
First area calculating means for calculating, as a first area, an area obtained by integrating a value obtained by subtracting the oxygen storage amount estimated value from an oxygen storage amount maximum value when the rear air-fuel ratio sensor output continues to be lean;
Second area calculating means for calculating, as a second area, an area obtained by integrating a value obtained by subtracting the oxygen storage amount minimum value from the oxygen storage amount estimated value when the rear air-fuel ratio sensor output continues to be rich;
Determining means for determining whether the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the first area and the second area;
A lean-side correction unit that corrects the stoichiometric point of the front air-fuel ratio sensor to a lean side when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the determination result. Fuel ratio control device.   2. The internal combustion engine according to claim 1, wherein the determination unit determines that the stoichiometric point of the front air-fuel ratio sensor is lean-shifted when a ratio between the first area and the second area exceeds a certain value. Engine air-fuel ratio control device. In a catalyst capable of storing or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value,
A front air-fuel ratio sensor installed upstream of the catalyst;
A rear air-fuel ratio sensor installed downstream of the catalyst;
Air-fuel ratio control means for estimating an oxygen storage amount based on the output of the front air-fuel ratio sensor and controlling the air-fuel ratio so that the oxygen storage amount estimated value matches a target value;
First length calculating means for calculating, as a first length, a value obtained by subtracting the oxygen storage amount estimated value from the oxygen storage amount maximum value when the rear air-fuel ratio sensor output crosses richer than stoichiometry;
A second length calculating means for calculating, as a second length, a value obtained by subtracting the oxygen storage amount minimum value from the oxygen storage amount estimated value when the rear air-fuel ratio sensor output crosses leaner than stoichiometric;
Determining means for determining whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the first length and the second length;
A lean-side correction unit that corrects the stoichiometric point of the front air-fuel ratio sensor to a lean side when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the determination result. Fuel ratio control device.   The said determination means determines that the stoichiometric point of the said front air fuel ratio sensor is lean-shifting when ratio of said 1st length and 2nd length exceeds a fixed value. An air-fuel ratio control apparatus for an internal combustion engine. In a catalyst capable of storing or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value,
A front air-fuel ratio sensor installed upstream of the catalyst;
A rear air-fuel ratio sensor installed downstream of the catalyst;
Air-fuel ratio control means for estimating an oxygen storage amount based on the output of the front air-fuel ratio sensor and controlling the air-fuel ratio so that the oxygen storage amount estimated value matches a target value;
A third area calculating means for calculating, as a third area, an area obtained by integrating the lean side output when the rear air-fuel ratio sensor output continues to be lean;
A fourth area calculating means for calculating, as a fourth area, an area obtained by integrating the rich side output when the rear air-fuel ratio sensor output continues to be rich;
Determining means for determining whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the third area and the fourth area;
A lean-side correction unit that corrects the stoichiometric point of the front air-fuel ratio sensor to a lean side when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the determination result. Fuel ratio control device.   The internal combustion engine according to claim 5, wherein the determination unit determines that the stoichiometric point of the front air-fuel ratio sensor is lean-shifted when a ratio of the third area and the fourth area exceeds a certain value. Engine air-fuel ratio control device. In a catalyst capable of storing or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value,
A front air-fuel ratio sensor installed upstream of the catalyst;
A rear air-fuel ratio sensor installed downstream of the catalyst;
Air-fuel ratio control means for estimating an oxygen storage amount based on the output of the front air-fuel ratio sensor and controlling the air-fuel ratio so that the oxygen storage amount estimated value matches a target value;
Amplitude number calculating means for calculating the number of times the rear air-fuel ratio sensor output swings from the minimum value to the maximum value or from the maximum value to the minimum value per fixed interval;
Determining means for determining whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the number of amplitudes per fixed section;
A lean-side correction unit that corrects the stoichiometric point of the front air-fuel ratio sensor to a lean side when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the determination result. Fuel ratio control device.   8. The internal combustion engine air conditioner according to claim 7, wherein the determination means determines that the stoichiometric point of the front air-fuel ratio sensor is lean-shifted when the number of amplitudes per fixed section exceeds a predetermined number. Fuel ratio control device. In a catalyst capable of storing or releasing oxygen in the exhaust gas according to the exhaust air-fuel ratio, and an apparatus for controlling the air-fuel ratio so that the oxygen storage amount becomes a target value,
A front air-fuel ratio sensor installed upstream of the catalyst;
A rear air-fuel ratio sensor installed downstream of the catalyst;
Air-fuel ratio control means for estimating an oxygen storage amount based on the output of the front air-fuel ratio sensor and controlling the air-fuel ratio so that the oxygen storage amount estimated value matches a target value;
A rich slice level at which the rear air-fuel ratio sensor output does not cross when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted;
The lean side slice level that the rear air-fuel ratio sensor output crosses when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted,
Continuous crossing number calculating means for calculating the number of times that the rear air-fuel ratio sensor output crosses the lean side slice level continuously in the same direction without crossing the rich side slice level per fixed interval;
Determining means for determining whether or not the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the number of continuous crossings per fixed section;
A lean-side correction unit that corrects the stoichiometric point of the front air-fuel ratio sensor to a lean side when the stoichiometric point of the front air-fuel ratio sensor is lean-shifted based on the determination result. Fuel ratio control device.   10. The internal combustion engine according to claim 9, wherein the determination unit determines that the stoichiometric point of the front air-fuel ratio sensor is lean-shifted when the number of continuous crossings per predetermined section exceeds a predetermined number. 10. Air-fuel ratio control device. A determination means for determining whether or not the driving condition is continuously staying in a predetermined learning area;
9. The air-fuel ratio control apparatus for an internal combustion engine according to claim 7 or 8, wherein the number of times of the amplitude is calculated when an operation condition continues to stay in the learning region based on the determination result. A determination means for determining whether or not the driving condition is continuously staying in a predetermined learning area;
The air-fuel ratio control apparatus for an internal combustion engine according to claim 9 or 10, wherein the number of times of continuous crossing is calculated when an operation condition continues to stay in the learning region based on the determination result.   The air-fuel ratio control apparatus for an internal combustion engine according to claim 11 or 12, wherein the learning region is a region excluding a high air flow region. The lean side correction means is
Air-fuel ratio conversion means for converting the output of the front air-fuel ratio sensor into an air-fuel ratio;
Shift correction amount calculation means for calculating a shift correction amount of a positive value when it is determined that the stoichiometric point of the front air-fuel ratio sensor is lean-shifted;
11. An actual air / fuel ratio calculating means for calculating, as an actual air / fuel ratio, a value obtained by adding the shift correction amount to a value obtained by converting the output of the front air / fuel ratio sensor into an air / fuel ratio. The air-fuel ratio control apparatus for an internal combustion engine according to any one of the above. The air-fuel ratio control means includes
Deviation calculating means for calculating a deviation between the oxygen storage amount estimated value and the target value;
Target air-fuel ratio calculating means for calculating the target air-fuel ratio based on this deviation;
Air-fuel ratio feedback correction coefficient calculating means for calculating an air-fuel ratio feedback correction coefficient so that the actual air-fuel ratio matches the target air-fuel ratio;
15. The air-fuel ratio control device for an internal combustion engine according to claim 14, comprising fuel supply amount correction means for correcting the fuel supply amount at the time of stoichiometry with the air-fuel ratio feedback correction coefficient.   16. The air-fuel ratio control apparatus for an internal combustion engine according to claim 14, wherein the shift correction amount is a learned value.

Images