JP4270073B2 - Air-fuel ratio control device for internal combustion engine - Google Patents

Description

  The present invention relates to an air-fuel ratio control apparatus for an internal combustion engine, and more particularly to an air-fuel ratio control apparatus suitable as an apparatus for controlling the air-fuel ratio of an on-vehicle internal combustion engine.

2. Description of the Related Art Conventionally, as disclosed in, for example, Japanese Patent Application Laid-Open No. 7-197837, an internal combustion engine is known that includes two exhaust gas sensors in an exhaust passage of the internal combustion engine. This internal combustion engine includes an air-fuel ratio sensor (a sensor that exhibits linear characteristics with respect to the air-fuel ratio) upstream of a catalyst disposed in the exhaust passage, and an O ₂ sensor (so-called Z with respect to the air-fuel ratio) downstream of the catalyst. Sensor).

In the conventional internal combustion engine, main feedback control is executed based on the output of the upstream air-fuel ratio sensor, while sub feedback control is executed based on the output of the downstream O ₂ sensor. In the main feedback control, the fuel injection amount is controlled so that the air-fuel ratio of the exhaust gas flowing into the catalyst matches the control target air-fuel ratio. In the sub-feedback control, more specifically, the output of the O ₂ sensor disposed downstream of the catalyst is stoichiometric output so that the air-fuel ratio of the exhaust gas flowing out downstream of the catalyst becomes the stoichiometric air-fuel ratio. Thus, the content of the main feedback control is corrected. According to these controls, the air-fuel ratio downstream of the catalyst can be systematically maintained at a value close to the theoretical air-fuel ratio, and excellent emission characteristics can be realized.

More specifically, in the sub-feedback control executed in the conventional internal combustion engine, the correction amount ΔV _{A / F} for correcting the content of the main feedback control is calculated by PID control as follows. .
ΔV _{A / F} = KP ・ ΔV _O2 + KI ・ (SUMΔV _O2 ) + KD ・ (dΔV _O2 )
However, in the above formulas, [Delta] V _O2 output deviation between the output V _O2 and the reference output V _O2 s the downstream O ₂ sensor _{(ΔV O2 = V O2 -V O2} s), SUMΔV O2 is the output deviation [Delta] V _O2 The integrated value, dΔV _O2 is the differential value of the output deviation ΔV _O2 , and KP, KI, KD are the proportional term gain, integral term gain, and differential term gain, respectively.

JP-A-7-197837 JP-A-8-177568 JP-A-52-144536

  By the way, in an internal combustion engine, it is possible to intentionally control the air-fuel ratio downstream of the catalyst to an air-fuel ratio that is slightly deviated to the rich side or lean side with respect to the theoretical air-fuel ratio in accordance with the operating state, the characteristics of the catalyst, and the like. It may be desirable. In such a case, a method of changing the value of the proportional term gain KP to a different value on the rich side and the lean side has been used in the internal combustion engine that executes the sub-feedback control as in the conventional technique. . If the configuration of the above prior art is only P control, it is possible to control to an air-fuel ratio with a target bias by this method.

However, when the I control is performed in addition to the P control, the deviation from the control target air-fuel ratio (the output deviation ΔV _O2 of the O ₂ sensor) is accumulated in the integral term KI · (SUMΔV _O2 ). As a result, transiently, the control target air-fuel ratio can be biased to either the rich side or the lean side by changing the gain of P control, which has a faster response speed than I control. Since the I control acts so as to cancel the bias of the air-fuel ratio by the P control (that is, by the setting of the proportional term gain KP), the air-fuel ratio downstream of the catalyst cannot be biased to the target air-fuel ratio. .

  The present invention has been made to solve the above-described problems. The air-fuel ratio downstream of the catalyst is reduced to the stoichiometric air-fuel ratio while suppressing the intentional air-fuel ratio bias caused by the P control from being canceled by the I control. An object of the present invention is to provide a control device for an internal combustion engine that can be controlled to an air-fuel ratio with a desired bias.

In order to achieve the above object, a first invention provides a catalyst disposed in an exhaust passage of an internal combustion engine,
An upstream exhaust gas sensor disposed upstream of the catalyst;
A downstream side exhaust gas sensor disposed downstream of the catalyst;
Main feedback means for correcting the fuel injection amount based on the output of the upstream side exhaust gas sensor so that the air-fuel ratio upstream of the catalyst becomes the target air-fuel ratio;
Sub-feedback means for correcting the correction by the main feedback means so that the air-fuel ratio downstream of the catalyst becomes the stoichiometric air-fuel ratio based on the output of the downstream exhaust gas sensor,
The sub-feedback means includes
When the downstream exhaust gas sensor emits a rich output, a value obtained by multiplying an output deviation between the output of the downstream exhaust gas sensor and a target value of the output by a rich time proportional term gain is calculated as a proportional term, When the downstream exhaust gas sensor emits a lean output, a value obtained by multiplying the output deviation by a lean time proportional term gain is calculated as a proportional term, and the lean time proportional term gain is calculated as the rich time proportional term. A proportional term calculation means for setting the gain larger than the gain;
Integration term calculation means for calculating an output deviation integrated value by an integration process of integrating a value obtained by multiplying the output deviation by an integration gain every predetermined processing cycle, and calculating an integral term proportional to the output deviation integrated value. ,
The integral term calculating means is configured to increase the integrated gain used when the downstream exhaust gas sensor emits a lean output compared to the integrated gain used when the downstream exhaust gas sensor emits a rich output. It is characterized by setting.

In order to achieve the above object, a second invention provides a catalyst disposed in an exhaust passage of an internal combustion engine,
An upstream exhaust gas sensor disposed upstream of the catalyst;
A downstream side exhaust gas sensor disposed downstream of the catalyst;
Main feedback means for correcting the fuel injection amount based on the output of the upstream side exhaust gas sensor so that the air-fuel ratio upstream of the catalyst becomes the target air-fuel ratio;
Sub-feedback means for correcting the correction by the main feedback means so that the air-fuel ratio downstream of the catalyst becomes the stoichiometric air-fuel ratio based on the output of the downstream exhaust gas sensor,
The sub-feedback means includes
When the downstream exhaust gas sensor emits a rich output, a value obtained by multiplying an output deviation between the output of the downstream exhaust gas sensor and a target value of the output by a rich time proportional term gain is calculated as a proportional term, When the downstream exhaust gas sensor emits a lean output, a value obtained by multiplying the output deviation by a lean time proportional term gain is calculated as a proportional term, and the rich time proportional term gain is calculated as the lean time proportional term. A proportional term calculation means for setting the gain larger than the gain;
Integration term calculation means for calculating an output deviation integrated value by an integration process of integrating a value obtained by multiplying the output deviation by an integration gain every predetermined processing cycle, and calculating an integral term proportional to the output deviation integrated value. ,
The integral term calculation means is configured to increase the integrated gain used when the downstream exhaust gas sensor emits a rich output as compared to the integrated gain used when the downstream exhaust gas sensor emits a lean output. It is characterized by setting.

  According to a third aspect of the present invention, in the first or second aspect of the invention, the integral term calculation means is rich in the integrated gain used when the downstream side exhaust gas sensor emits a lean output and the downstream side exhaust gas sensor. The integrated gain is set so that a ratio of the integrated gain used when output is generated corresponds to a ratio of the lean proportional term gain and the rich proportional term gain.

  According to the first aspect of the invention, the output of the downstream side exhaust gas sensor is controlled by the action of the proportional term calculation means so that it becomes longer in rich output than in lean output. Further, according to the present invention, when performing integration processing by the integral term calculation means, the output deviation integrated value calculated when the downstream side exhaust gas sensor indicates a rich output is compared with the rate of change for each processing cycle. Thus, the output deviation integrated value calculated when the downstream side exhaust gas sensor indicates a lean output is controlled so as to increase the rate of change for each processing cycle. For this reason, it is suppressed that the value of the integral term largely changes in the negative direction. Therefore, according to the present invention, the air-fuel ratio downstream of the catalyst is accurately controlled to the originally intended rich air-fuel ratio while effectively suppressing the intentional air-fuel ratio bias caused by the control from being canceled by the I control. It becomes possible.

  According to the second aspect of the invention, the output of the downstream side exhaust gas sensor is controlled by the action of the proportional term calculation means so as to be longer to the lean output than the rich output. Further, according to the present invention, when performing integration processing by the integral term calculation means, the output deviation integrated value calculated when the downstream side exhaust gas sensor indicates a lean output is compared with the rate of change for each processing cycle. Thus, the output deviation integrated value calculated when the downstream side exhaust gas sensor indicates a rich output is controlled so as to increase the rate of change for each processing cycle. For this reason, it is suppressed that the value of an integral term changes greatly in the positive direction. Therefore, according to the present invention, the air-fuel ratio downstream of the catalyst is accurately controlled to the originally intended lean air-fuel ratio while effectively suppressing the intentional air-fuel ratio bias caused by the control from being canceled by the I control. It becomes possible.

  According to the third invention, it is possible to prevent the value of the integral term from deviating greatly in one of the positive and negative directions. For this reason, according to the present invention, it is possible to control the air-fuel ratio downstream of the catalyst to the originally intended rich or lean air-fuel ratio with extremely high accuracy without canceling the intentional air-fuel ratio bias due to the control by the I control. It becomes.

Embodiment 1 FIG.
[Configuration of Device of Embodiment 1]
FIG. 1 is a diagram for explaining the configuration of an air-fuel ratio control apparatus according to Embodiment 1 of the present invention. As shown in FIG. 1, the apparatus of this embodiment includes an upstream catalyst (S / C) 12 and a downstream catalyst (U / F) 14 disposed in an exhaust passage 10 of the internal combustion engine. The upstream catalyst 12 and the downstream catalyst 14 are all three-way catalysts that can simultaneously purify CO, HC, and NOx.

A main air-fuel ratio sensor 16 and a sub O ₂ sensor 18 are disposed upstream and downstream of the upstream catalyst 12, respectively. The main air-fuel ratio sensor 16 is a sensor that emits a substantially linear output with respect to the air-fuel ratio A / F of the exhaust gas flowing into the upstream catalyst 12. On the other hand, the sub O ₂ sensor 18 generates a rich output (for example, 0.8 V) when the exhaust gas flowing out from the upstream catalyst 12 is rich with respect to the stoichiometric air-fuel ratio, and the exhaust gas is lean. In this case, the sensor generates a lean output (for example, 0.2 V).

The output of the main air-fuel ratio sensor 16 and the output of the sub O ₂ sensor 18 are respectively supplied to an ECU (Electronic Control Unit) 20. The ECU 20 is further connected to an air flow meter 22, a rotation speed sensor 24, a fuel injection valve 26, and the like. The air flow meter 22 is a sensor that detects an intake air amount Ga of the internal combustion engine. The rotational speed sensor 24 is a sensor that generates an output corresponding to the engine rotational speed Ne. The fuel injection valve 26 is an electromagnetic valve for injecting fuel into the intake port of the internal combustion engine.

[Air-fuel ratio control in Embodiment 1]
FIG. 2 is a control block diagram for explaining the contents of the air-fuel ratio control executed in the air-fuel ratio control apparatus of this embodiment. In FIG. 2, Cat12 corresponds to the upstream catalyst 12 shown in FIG. Further, A / Fs 16 and Vox 18 in FIG. ₂ correspond to the main air-fuel ratio sensor 16 and the sub O ₂ sensor 18 shown in FIG. 1, respectively. Further, Eng30 in FIG. 2 means the main body of the internal combustion engine in which the system shown in FIG. 1 is incorporated.

  In the system of the present embodiment, the main air-fuel ratio sensor 16 generates an output A / Fs [v] corresponding to the air-fuel ratio of the exhaust gas that flows out from the internal combustion engine 30 and flows into the upstream catalyst 12. Specifically, the output A / Fs becomes larger as the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 is larger, that is, as the air-fuel ratio is leaner.

The sub O ₂ sensor 18 generates an output Vox [v] corresponding to the oxygen concentration in the exhaust gas flowing from the upstream catalyst 12 toward the downstream catalyst 14. The output Vox [v] of the sub O ₂ sensor 18 is supplied to the subtractor 32. The differencer 32 calculates an output deviation Dvox = Voxref−Vox between the output Vox and the target voltage Voxref of the sub-feedback. Then, the sub-feedback controller 34 determines the “proportional term Dvox · GainP” with respect to the output deviation Dvox and the output change amount Dlvox of the sub O ₂ sensor 18 (= previous output Vox (i−1) −current Vox (i )) For “derivative term Dlvox · GainD”. Further, the sub-feedback controller 34 calculates “integral term Sumvox · GainI” with respect to the output deviation Dvox. More specifically, the output deviation integrated value Sumvox in the integral term Sumvox · GainI adds the previous output deviation integrated value Sumvox (i−1) to the value obtained by multiplying the current output deviation Dvox by an integrated gain Kgain described later. Thus, Sumvox (i) = Sumvox (i−1) + (Dvox · Kgain) is calculated.

  The sub feedback controller 34 calculates the sum “Dvox · GainP + Dlvox · GainD + Sumvox · GainI” of the above terms as the sub feedback correction amount Vsfb [v]. This sub feedback correction amount Vsfb is added to the output A / Fs of the main air-fuel ratio sensor 16 in the adder 36.

  In the control block shown in FIG. 2, a low-pass filter 38 and an integrator 40 are inserted between the sub feedback controller 34 and the adder 36. The low-pass filter 38 is a block for extracting a change occurring in the integral term Sumvox · GainI calculated by the sub-feedback controller 34. The accumulator 40 is a block for accumulating the change and storing it as a sub-feedback learning value SFBG. The adder 36 calculates the correction voltage value Vfb by adding the sub feedback learning value SFBG to the output A / Fs and the sub feedback correction amount Vsfb.

  Further, in this apparatus, when the component extracted by the low-pass filter 38 is taken into the sub-feedback learning value SFBG in the accumulator 40, the sub-feedback controller 34 uses the same value that is taken there as the integral term Sumvox · GainI. It is going to be reduced from. According to these processes, while the integral term Sumvox · GainI itself is maintained in the vicinity of the reference value, the increase / decrease occurring in the integral term Sumvox · GainI can be extracted as the learning value SFBG and supplied to the adder 36.

  The corrected voltage value Vfb is supplied to the voltage-air fuel ratio converter 42. The voltage-air-fuel ratio converter 42 stores a voltage-air-fuel ratio conversion map, and determines a corrected air-fuel ratio eabyf corresponding to the correction voltage value Vfb according to the map. More specifically, according to the processing of the voltage-air-fuel ratio converter 42, the larger the correction voltage value Vfb, the larger the correction air-fuel ratio eabyf, that is, a leaner value.

  The ECU 20 executes main feedback control based on the corrected air-fuel ratio eabyf determined as described above. Specifically, according to the main feedback control, increase / decrease correction is performed on the base fuel injection amount so that the deviation between the corrected air-fuel ratio eabyf and the target air-fuel ratio (theoretical air-fuel ratio) disappears. As a result, the final fuel injection amount is determined so that the corrected air-fuel ratio eabyf changes to the stoichiometric air-fuel ratio. Note that the content of the main feedback control is not the main part of the present invention and is already a known matter, and therefore, further explanation is omitted here. Hereinafter, correcting the content of the main feedback control by reflecting the sub feedback correction amount Vsfb and the sub feedback learning value SFBG in the output A / Fs is referred to as “sub feedback control”.

[Characteristics of sub feedback correction amount Vsfb and sub feedback learning]
Next, the characteristic of the correction voltage value Vfb, more specifically, the characteristic of the sub feedback correction amount Vsfb will be described. As described above, the correction voltage value Vfb is obtained by adding the sub-feedback correction amount Vsfb = Dvox · GainP + Dlvox · GainD + Sumvox · GainI and the sub-feedback learning value SFBG to the output A / Fs of the main air-fuel ratio sensor 16. Hereinafter, characteristics of each term on the right side of the sub feedback correction amount Vsfb and sub feedback learning will be described in detail.

(1) Characteristics of “proportional term Dvox · GainP” and “derivative term Dlvox · GainD” Here, for convenience of explanation, first, the correction voltage value Vfb is compared with the air-fuel ratio to be realized, that is, the stoichiometric air-fuel ratio. Assume that the value is excessive.

  The voltage-air-fuel ratio converter 42 determines the corrected air-fuel ratio eabyf based on the corrected voltage value Vfb as described above. When the corrected voltage value Vfb is an excessive value with respect to the stoichiometric air-fuel ratio, the corrected air-fuel ratio eabyf is determined to be a lean value as compared with the stoichiometric air-fuel ratio. In this case, in the main feedback control, fuel injection amount correction is performed so that the corrected air-fuel ratio eabyf becomes the stoichiometric air-fuel ratio. As a result, the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 is controlled to a value biased to the rich side as compared with the stoichiometric air-fuel ratio. When the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 is biased toward the rich side, the air-fuel ratio of the exhaust gas flowing out downstream of the upstream catalyst 12 is also biased toward the rich side.

When the air-fuel ratio of the exhaust gas flowing out from the upstream catalyst 12 changes in the vicinity of the theoretical air-fuel ratio, the sub O ₂ sensor 18 has a rich-side upper limit value of about 0.8 [v] and 0.2 [ The output Vox is sensitively changed between the lower limit of the lean side of about v]. For this reason, if the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 has shifted to the rich side, the output Vox of the sub O ₂ sensor 18 will detect the rich side upper limit (0.8) when the influence of the shift is detected. [v]) suddenly change.

In the subtractor 32, the median value (0.5 [v]) of the above-mentioned rich side upper limit value (0.8 [v]) and lean side lower limit value (0.2 [v]) is the sub-feedback target voltage Voxref. Used as Therefore, the output deviation Dvox calculated by the differentiator 32 is a value having a negative sign when the output Vox of the sub O ₂ sensor 18 is shifted to the rich side (0.8 [v] side). . In this case, the proportional term Dvox · GainP included in the sub feedback correction amount Vsfb is proportional to the rich shift amount of the output Vox of the sub O ₂ sensor 18 and has a negative sign. Similarly, the differential term Dlvox · GainD included in Vsfb has a value corresponding to the degree of sudden change in the output Vox and has a negative sign.

  As described above, the proportional term Dvox · GainP and the differential term Dlvox · GainD are added to the output A / Fs of the main air-fuel ratio sensor 16 in the adder 36, thereby forming the basis of the correction voltage value Vfb. If the proportional terms Dvox · GainP and the differential terms Dlvox · GainD are both negative values, the correction voltage value Vfb is corrected to a small value. More specifically, the correction voltage value Vfb is reduced to an appropriate value so that the rich shift downstream of the upstream catalyst 12 disappears.

  When the correction voltage value Vfb is reduced as described above, the correction air-fuel ratio eabyf is decreased by the amount corresponding to the decrease, and as a result, becomes a value corresponding to the theoretical air-fuel ratio. When the main feedback control is executed based on the corrected corrected air-fuel ratio eabyf, the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 becomes the stoichiometric air-fuel ratio. As such, it is appropriately leaned.

  The above description is based on the assumption that the correction voltage value Vfb is excessive compared to the stoichiometric air-fuel ratio, but the function for the sub-feedback correction amount Vsfb to correct the exhaust air-fuel ratio to a value close to the stoichiometric air-fuel ratio is The same occurs when the correction voltage value Vfb is too small compared to the stoichiometric air-fuel ratio. That is, in this case, first, the exhaust air-fuel ratio downstream of the upstream catalyst 12 is controlled to a lean value. When the influence appears downstream of the upstream catalyst 12, the proportional terms Dvox · GainP and the differential terms Dlvox · GainD included in the sub-feedback correction amount Vsfb both have a magnitude that cancels the lean shift and are positive. The value has a sign.

  If the proportional terms Dvox · GainP and the differential terms Dlvox · GainD are both positive values, the correction voltage value Vfb is corrected to a larger value. When the correction voltage value Vfb is corrected to a large value, the correction air-fuel ratio eabyf is also changed to a large value. When the main feedback control is executed based on the corrected corrected air-fuel ratio eabyf, the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 becomes the stoichiometric air-fuel ratio. As shown in FIG.

  As described above, according to the sub-feedback control using the sub-feedback correction amount Vsfb including the proportional terms Dvox · GainP and the differential terms Dlvox · GainD, the exhaust air-fuel ratio downstream of the upstream catalyst 12 is matched with the stoichiometric air-fuel ratio. Can do. More specifically, according to the proportional term Dvox · GainP, the magnitude of the output deviation Dvox can be directly reflected in the correction of the correction voltage value Vfb. Further, according to the differential term Dlvox · GainD, the correction voltage value Vfb can be corrected in a direction to cancel the tendency of the output change amount Dlvox.

  That is, according to the sub-feedback control using the sub-feedback correction amount Vsfb including the proportional terms Dvox · GainP and the differential terms Dlvox · GainD, when the exhaust air-fuel ratio downstream of the upstream catalyst 12 deviates from the stoichiometric air-fuel ratio, and When an attempt is made to deviate, the correction voltage value Vfb can be quickly corrected in a direction to eliminate the deviation and in a direction to cancel the change. For this reason, according to the system of the present embodiment, the air-fuel ratio downstream of the upstream catalyst 12 can be accurately maintained at a value close to the theoretical air-fuel ratio during execution of the main feedback control.

(2) Characteristics of “Integral Term Sumvox · GainI” Next, it is assumed that the fluctuation center of the correction voltage value Vfb is shifted in an excessive direction as compared with the theoretical air-fuel ratio. When the fluctuation center of the correction voltage value Vfb has shifted to an excessive value with respect to the theoretical air-fuel ratio, a value in which the fluctuation center of the exhaust air-fuel ratio upstream of the upstream catalyst 12 is richly biased by the function of the main feedback control. It becomes. In this case, the fluctuation center of the exhaust air-fuel ratio downstream of the upstream catalyst 12 is also a richly biased value. On the other hand, when the correction voltage value Vfb is shifted in an excessively small direction compared to the stoichiometric air-fuel ratio, the fluctuation center of the exhaust air-fuel ratio downstream of the upstream catalyst 12 becomes a value biased lean.

For this reason, when the air-fuel ratio of the exhaust gas flowing out from the upstream catalyst 12 is richly biased, the output Vox of the sub O ₂ sensor 18 is also biased to the rich side (0.8 [v] side). On the other hand, when the air-fuel ratio is biased toward lean, the output Vox is also biased toward the lean side (0.2 [v] side). The integral term Sumvox · GainI included in the sub-feedback correction amount Vsfb is obtained by multiplying the output deviation integrated value Sumvox based on the output deviation Dvox by a predetermined integral term gain GainI. When the center of fluctuation of the fuel ratio (hereinafter referred to as the “downstream air-fuel ratio fluctuation center”) is biased toward the rich side, a negative value is obtained. On the other hand, when the center of fluctuation of the downstream air-fuel ratio is biased toward the lean side, Positive value. The absolute value of the integral term Sumvox · GainI is a value corresponding to the amount of deviation of the downstream air-fuel ratio fluctuation center from the theoretical air-fuel ratio.

(When there is a rich shift at the downstream air-fuel ratio fluctuation center)
The integral term Sumvox · GainI included in the sub-feedback correction amount Vsfb is added to the output A / Fs of the main air-fuel ratio sensor 16 by the adder 36 to be the basis of the correction voltage value Vfb. When the downstream air-fuel ratio fluctuation center is biased to the rich side and the integral term Sumvox · GainI shows a negative value, the correction voltage value Vfb is corrected to a small value. More specifically, in this case, the correction voltage value Vfb is appropriately reduced so that the rich shift at the downstream air-fuel ratio fluctuation center disappears.

  When the correction voltage value Vfb is reduced as described above, the correction air-fuel ratio eabyf is decreased (enriched) by the amount corresponding to the decrease. Then, the main feedback control is executed based on the corrected corrected air-fuel ratio eabyf, whereby the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 is corrected in the lean direction. When the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 is corrected in the lean direction, the air-fuel ratio of the exhaust gas flowing out from the upstream catalyst 12 naturally shifts in the lean direction. As a result, the rich shift at the downstream air-fuel ratio fluctuation center is eliminated.

(When lean deviation occurs at the center of downstream air-fuel ratio fluctuation)
On the other hand, when the lean deviation occurs at the center of the downstream air-fuel ratio fluctuation, and as a result, the integral term Sumvox · GainI shows a positive value, the adder 36 corrects the correction voltage value Vfb to a large value. . That is, the correction voltage value Vfb is increased moderately to such an extent that the lean shift at the downstream air-fuel ratio fluctuation center disappears. When the correction voltage value Vfb is increased as described above, the correction air-fuel ratio eabyf is increased (leaned) by a value corresponding to the increase. When the main feedback control is executed based on the corrected air-fuel ratio eabyf after the change, the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 is corrected in the rich direction, and as a result, the lean air-fuel ratio fluctuation center lean The deviation is eliminated.

(Effects of using the integral term Sumvox / GainI)
Integral term Sumvox · GainI that the output Vox of the sub O ₂ sensor 18 and the underlying, regardless of whether the output Vox of the sub O ₂ sensor 18 indicates any value as the instantaneous value, the exhaust gas in the downstream of the upstream catalyst 12 This shows how the air-fuel ratio has a deviation from the stoichiometric air-fuel ratio. For this reason, according to the sub-feedback control using the sub-feedback correction amount Vsfb including the integral terms Sumvox · GainI, the empty air-fuel ratio of the exhaust gas flowing out from the upstream catalyst 12 is not affected by the instantaneous air-fuel ratio. Correction for converging the fluctuation center of the fuel ratio to the stoichiometric air-fuel ratio can be realized.

(3) Explanation of Sub-Feedback Learning The sub-feedback correction amount Vsfb described above is based on the air-fuel ratio of the exhaust gas flowing into the upstream catalyst 12 for making the exhaust air-fuel ratio downstream of the upstream catalyst 12 stoichiometric and the actual upstream air-fuel ratio. This value corresponds to the deviation from the fuel ratio. The average value of the integral term Sumvox · GainI included in the sub-feedback correction amount Vsfb corresponds to a constant deviation inherent in the air-fuel ratio control (deviation caused by individual differences or changes with time of the internal combustion engine 30). ing.

  If the average value of the integral terms Sumvox · GainI is transferred to the sub-feedback learning value SFBG, the permanent deviation inherent in the system can be canceled by the sub-feedback learning value SFBG. In this case, the role of the integral term Sumvox · GainI is only to absorb temporary deviations caused by various causes. If the integral term Sumvox · GainI is also responsible for the absorption of constant deviations in the system, it takes a long time to converge. On the other hand, if the role is only to absorb temporary deviation, the time required for the convergence can be sufficiently shortened. For this reason, transferring a part of the integral term Sumvox · GainI to the sub-feedback learning value SFBG is extremely useful for shortening the control convergence time.

[Features of Embodiment 1]
Next, features of the first embodiment of the present invention will be described with reference to FIG. 3 and FIG.
In the internal combustion engine 30, the air-fuel ratio of the exhaust gas flowing out downstream of the upstream catalyst 12 is intentionally set slightly to the rich side or lean side with respect to the theoretical air-fuel ratio in accordance with the operating state, the characteristics of the catalyst, and the like. It may be desirable to control the air-fuel ratio to be biased, that is, to control the output of the sub O ₂ sensor 18 to be slightly biased to the rich side or the lean side. Hereinafter, in the present specification, controlling the air-fuel ratio biased to the rich side with respect to the theoretical air-fuel ratio is referred to as “rich approach”, and similarly controlling to the air-fuel ratio biased to the rich side is referred to as “ Sometimes referred to as “lean alignment”. Here, it is assumed that the explanation is advanced by taking rich as an example.

(Conventional rich shift control)
First, with reference to FIG. 3, a rich approach technique performed in a conventional air-fuel ratio control apparatus and adverse effects caused by the technique will be described. In the conventional air-fuel ratio control device here, it is assumed that sub-feedback control including P control and I control is performed as in the above-described device of the present embodiment. 3A shows a waveform of the output Vox of the sub O ₂ sensor 18, and FIG. 3B shows an integrated value corresponding to the output vox of the sub O ₂ sensor 18 of FIG. The waveform of the term Sumvox · GainI is shown.

The above-described conventional apparatus performs targeted rich shift control by setting the proportional term gain GainP included in the sub feedback correction amount Vsfb to be different between the rich side and the lean side. FIG. 3A shows a state in which the sub O ₂ output Vox is maintained on the rich side longer than the lean side by setting the proportional term gain GainP to such a gain. More specifically, in the above conventional technique, compared to rich when the proportional term gain GainPr used when the sub-O ₂ sensor 18 is emitting a rich output, the sub-O ₂ sensor 18 is emitting a lean output The lean-time proportional term gain GainPl used in this case is set large. According to such a setting, when the sub O ₂ output Vox shows a lean output, when the sub O ₂ output Vox shows a rich output in contrast to the effect that the P control cancels the lean deviation, , P control can weaken the effect of canceling the rich shift. Therefore, according to the conventional method, as shown in FIG. 3A, the sub O ₂ output Vox is maintained longer on the rich side than on the lean side.

In the conventional apparatus, as described above, sub-feedback control including I control is performed in addition to P control. Therefore, when the sub O ₂ output Vox is maintained longer on the rich side than on the lean side, the output deviation Dvox is simply calculated for each processing cycle in the integral term Sumvox · GainI included in the sub feedback correction amount Vsfb. When integrated, the waveform of the output deviation integrated value Sumvox, that is, the waveform of the integral term Sumvox · GainI obtained by multiplying the output deviation integrated value Sumvox by a certain integral term gain GainI is as shown in FIG. It looks like a curve. This is because when the sub O ₂ output Vox is producing a lean output, the output deviation Dvox is positive, so the output deviation integrated value Sumvox tends to increase, while the sub O ₂ output Vox produces a rich output. This is because the output deviation integrated value Sumvox tends to decrease because the output deviation Dvox is negative.

As described above, when the sub O ₂ output Vox is maintained longer on the rich side than on the lean side, as shown in FIG. 3B, the value of the integral term Sumvox · GainI changes greatly in the negative direction. As a result, lean correction is strongly promoted as the integral term Sumvox · GainI. In other words, according to the conventional method described above, by making the proportional term gain GainP different between the rich side and the lean side, the rich adjustment is made transiently by changing the gain of the P control, which has a faster response speed than the I control. Although it can be realized, the output deviation Dvox is accumulated in the integral term Sumvox · GainI, and therefore, the rich approach due to the action of the P control is canceled by the action of the I control. When the conventional apparatus performs the sub-feedback learning process as in the apparatus of the present embodiment, as shown in FIG. 3B, the value of the integral term Sumvox · GainI that has greatly shifted in the negative direction As a result, when the sub-feedback control is stopped, the lean value may be controlled instead.

When rich gathering is performed by the conventional method, as described above, the sub O ₂ output Vox is corrected so that the rich gathering is canceled by the action of the I control. For this reason, in a situation where the integral term Sumvox · GainI changes in the negative direction, the sub-O ₂ output Vox is actually a time during which the sub-O ₂ output Vox is maintained rich (hereinafter referred to as “rich maintenance time”). As shown in FIG. 3A, the ratio of the time maintained in lean (hereinafter referred to as “lean maintenance time”) does not repeat reversal at a constant interval, and eventually the rich maintenance time and lean The maintenance time is controlled to be equal. However, here, in order to easily explain the change indicated by the integral term Sumvox · GainI when the rich approach is performed by the conventional method, in FIG. 3A, the ratio between the rich maintenance time and the lean maintenance time is shown. Is shown to remain constant.

(Rich adjustment control of this embodiment)
FIG. 4 is a diagram for explaining the effect of the rich shift control used in the present embodiment. More specifically, FIG. 4A shows the waveform of the output Vox of the sub O ₂ sensor 18, and FIG. 4B corresponds to the output vox of the sub O ₂ sensor 18 of FIG. The waveform of the integral value, that is, the integral term Sumvox · GainI is shown. Also in the rich shift control method used in the present embodiment, the proportional term gain GainP included in the sub feedback correction amount Vsfb is set so that the rich time proportional term gain GainPr <the lean time proportional term gain GainPl is satisfied. In this case, the rich shift control is performed.

When such rich shift control is executed, the output for each processing cycle is performed regardless of whether the sub O ₂ output Vox indicates rich output or lean output as in the conventional method described above. If the deviation Dvox (i) is simply added to the previous output deviation integrated value Sumvox (i−1) with the same weight, the integral term Sumvox · GainI is negative because the rich maintenance time differs from the lean maintenance time. The value is greatly shifted (see FIG. 3B). Therefore, in the method of the present embodiment, the sub-O ₂ output Vox is calculated when performing the integration process of the output deviation integrated value Sumvox in order to prevent the output deviation integrated value Sumvox from greatly changing to a negative value during the rich shift control. Output deviation integrated value Sumvox calculated when sub-O ₂ output Vox indicates lean output with respect to the rate of change for each processing cycle of output deviation integrated value Sumvox calculated when indicates rich output The rate of change for each cycle was increased. The method will be specifically described below.

  As described above, in this embodiment, the output deviation integrated value Sumvox is set as Sumvox = Sumvox (i−1) + Dvox · Kgain, and the output deviation Dvox is not simply integrated, but the integration set as follows: The gain Kgain is multiplied by the output deviation Dvox. This integrated gain Kgain is a factor governing the slope of the curve of the integral term Sumvox · GainI in FIG. 4B (that is, the rate of change of the output deviation integrated value Sumvox).

Therefore, in this embodiment, the value of the integrated gain Kgain is set to the integrated gain used when the sub O ₂ output Vox shows a lean output from the integrated gain Kgain used when the sub O ₂ output Vox shows a rich output. The gain Kgain was set large. According to such a setting, the sub O ₂ output Vox is lean output compared to the rate of change of the output deviation integrated value Sumvox calculated for each processing cycle when the sub O ₂ output Vox indicates rich output. , The rate of change of the output deviation integrated value Sumvox calculated for each processing cycle can be increased, and the integral term Sumvox · GainI can be prevented from greatly changing in the negative direction. Furthermore, in this embodiment, the integrated gain Kgain is set as follows in order to surely prevent the integral term Sumvox · GainI from shifting in one direction (here, the negative direction because it is rich).

As described above, when the proportional term gain GainP is set so that the rich proportional term gain GainPr <the lean proportional term gain GainPl, if the sub O ₂ output Vox has a rich shift, the lean The shift is canceled more slowly than when the shift occurs. That is, due to the action of the proportional term gain GainP, the rich maintenance time becomes longer than the lean maintenance time as shown in FIG. Therefore, the ratio between the rich maintenance time and the lean maintenance time can be expressed as a ratio of the lean time proportional term gain GainPl and the rich time proportional term gain GainPr as shown in the following equation (1).
Rich maintenance time: Lean maintenance time = GainPl: GainPr (1)

  As described above, the integrated gain Kgain is a factor that governs the slope of the curve of the integral term Sumvox · GainI in FIG. Therefore, if the accumulated gain Kgain is multiplied by the rich maintenance time and the lean maintenance time set in the relationship of the above equation (1), the product of these changes is the change in the output deviation integrated value Sumvox during the rich maintenance time. This means the ratio between the amount and the change amount of the output deviation integrated value Sumvox during the lean maintenance time.

Therefore, in this embodiment, the ratio between the accumulated gain Kgain used in the case of rich output and the accumulated gain Kgain used in the case of lean output is expressed by the following formula (2): rich maintenance time and lean maintenance time The ratio is set to a reciprocal relationship with respect to the ratio (GainPl: GainPr).
Kgain (rich): Kgain (lean) = (1 / GainP l) :( 1 / GainP r)
= GainP r : GainP l (2)

  According to the integrated gain Kgain defined by the above equation (2), the change amount of the output deviation integrated value Sumvox during the rich maintenance time can be matched with the change amount of the output deviation integrated value Sumvox during the lean maintenance time. . That is, according to the setting of the integrated gain Kgain, the value of the integral term Sumvox · GainI varies between positive and negative as shown in FIG. 4B with a certain amplitude, and the integral term Sumvox · GainI It is possible to reliably prevent the value from deviating greatly in the negative direction. At this time, even if the fluctuation center of the output deviation integrated value Sumvox is deviated from zero in the initial stage of accumulation, the deviation of the fluctuation center is taken into the sub-feedback learning value SFBG. . As a result, the value of the integral term Sumvox · GainI is surely converged to zero. For this reason, according to the rich approach control method of the present embodiment, it is possible to realize the rich approach by the P control without being canceled by the action of the I control.

[Specific Processing in Embodiment 1]
FIG. 5 is a flowchart of a routine executed by the ECU 20 in the present embodiment in order to realize the above function. This routine is periodically executed at predetermined processing cycles such as every fixed crank. In the routine shown in FIG. 5, first, the output deviation Dvox of the sub O ₂ sensor 18 is calculated (step 100). The output deviation Dvox is calculated as Dvox = Voxref−Vox by subtracting the output Vox of the sub O ₂ sensor 18 from the target voltage Voxref of the sub feedback.

Next, the output change amount Dlvox of the sub O ₂ sensor 18 is calculated (step 102). The output change amount Dlvox is calculated as Dlvox = Vox (i−1) −Vox (i) by subtracting the output Vox (i) of the current processing cycle from the output Vox (i−1) of the previous processing cycle. .

  Next, the proportional term (P term) gain GainP included in the sub feedback correction amount Vsfb, that is, the rich time proportional term gain GainPr and the lean time proportional term gain GainPl are read (step 104). The ECU 20 stores a gain set so that the rich time proportional term gain GainPr <the lean time proportional term gain GainPl in order to enable the above-described rich shifting. Next, the integral term (I term) gain GainI included in the sub feedback correction amount Vsfb is read (step 106), and the differential term (D term) gain GainD is read (step 108).

Next, it is determined whether or not lean output is in progress (target voltage Voxref> output Vox) (step 110). As a result, when it is determined that the target voltage Voxref> the output Vox is not satisfied, that is, the sub O ₂ sensor 18 is generating a rich output, the integrated gain (P-term gain ratio) Kgain used in the current processing cycle. Is calculated (step 112). The ECU 20 stores an integrated gain Kgain set so as to satisfy the relationship of the above expression (2). That is, in this step 112, the integrated gain Kgain stored in the ECU 20 is referred to, and Kgain = 1.0 is set as the integrated gain Kgain used when the sub O ₂ sensor 18 emits a rich output. Next, when it is determined by the processing of step 110 that the sub O ₂ sensor 18 is producing a rich output, the rich time proportional term gain GainPr is selected as the proportional term gain GainP (step 114).

On the other hand, if it is determined in step 110 that the target voltage Voxref> the output Vox is satisfied, that is, the sub O ₂ sensor 18 is emitting a lean output, the integrated gain Kgain used in the current processing cycle is According to the relationship of the above equation (2), Kgain = (GainPl / GainPr) is calculated (step 116). Next, if it is determined by the processing of step 110 that the sub O ₂ sensor 18 is generating a lean output, the lean proportional gain GainPl is selected as the proportional term gain GainP (step 118).

  Next, an integration process of the output deviation Dvox is executed (step 120). Specifically, by this integration process, the output deviation integrated value Sumvox is calculated as Sumvox = Sumvox (i−1) + (Dvox · Kgain). Next, the sub feedback correction amount Vsfb is calculated using the elements obtained by the above processes (step 122). The sub feedback correction amount Vsfb is calculated as the sum of a proportional term, an integral term, and a differential term, that is, Vsfb = (Dvox · GainP) + (Sumvox · GainI) + (Dlvox · GainD).

  Next, the sub-feedback learning process is executed based on the sub-feedback correction amount Vsfb calculated in step 122 (step 124). Specifically, in the sub-feedback learning process, a process of calculating the low-pass value of the integral term Sumvox · GainI calculated in the above step 122 is performed, thereby extracting a change occurring in the integral term Sumvox · GainI. The Then, the learning value SFBG is updated with the value obtained by integrating the extracted low-pass values (changes in the integral terms Sumvox · GainI), that is, the steady values of the integral terms Sumvox · GainI are taken into the SFBG and taken in Subtract or add from the integral term Sumvox · GainI by the calculated value.

  According to the above processing, it is possible to control the air-fuel ratio downstream of the upstream catalyst 12 to the originally intended rich air-fuel ratio with high accuracy without canceling the rich approach by the P control by the I control. Further, the above processing shows the case of rich gathering, but lean gathering is performed by the same method as the above-described rich gathering except that the rich time proportional term gain GainPr> lean time proportional term gain GainPl is set. Can be realized.

Incidentally, in the first embodiment described above, the sensor disposed upstream of the upstream catalyst 12 is an air-fuel ratio sensor, and the sensor disposed downstream of the upstream catalyst 12 is an O ₂ sensor. Sensors that can be used in the present invention are not limited to these combinations. That is, any of these sensors may be any sensors that detect the characteristics of exhaust gas as a basis for feedback control, such as an air-fuel ratio sensor, an O ₂ sensor, a NOx sensor, and an HC sensor. In the first embodiment described above, sub-feedback control based on PID control is performed, but the present invention is naturally applicable to sub-feedback control based on PI control.

In the first embodiment described above, the upstream catalyst 12 is the “catalyst” in the first invention, the main air-fuel ratio sensor 16 is the “upstream exhaust gas sensor” in the first invention, and the sub O ₂ sensor. Reference numeral 18 corresponds to the “downstream exhaust gas sensor” in the first aspect of the invention. The ECU 20 executes the main feedback control based on the corrected air-fuel ratio eabyf obtained from the corrected voltage value Vfb calculated based on the output A / Fs, the sub-feedback correction amount Vsfb, and the sub-feedback learning value SFBG. The “main feedback means” according to the first aspect of the invention causes the ECU 20 to execute the processing of the routine shown in FIG. The “proportional term calculation means” in the first invention is executed by executing the processing of steps 122, 122, and the “integral term” in the first invention is executed by the ECU 20 executing the processing of steps 100, 106, 110-122. "Calculation means" is realized respectively.

It is a figure for demonstrating the structure of the air fuel ratio control apparatus of Embodiment 1 of this invention. It is a control block diagram for demonstrating the content of the air fuel ratio control performed in the air fuel ratio control apparatus of Embodiment 1 of this invention. It is a figure for demonstrating the technique of the rich approach currently performed with the conventional air fuel ratio control apparatus, and the harmful effect which the technique brings. It is a figure for demonstrating the effect by the rich shift control used in Embodiment 1 of this invention. It is a flowchart of the routine performed in Embodiment 1 of the present invention.

Explanation of symbols

10 Exhaust passage 12 Upstream catalyst 16 Main air-fuel ratio sensor 18 Sub O ₂ sensor 20 ECU (Electronic Control Unit)
30 Internal combustion engine 32 Subtractor 34 Sub feedback controller 36 Adder 38 Low pass filter 40 Accumulator
Vfb correction voltage value
Output of A / Fs main air-fuel ratio sensor
Vsfb Sub feedback correction amount
SFBG sub feedback learning value
Vox sub O ₂ sensor output
Voxref Sub feedback target voltage
Dvox output deviation
Dlvox output change
Proportional term based on output of Dvox / GainP sub-O ₂ sensor
Dlvox · GainD Sub O ₂ sensor based differential term
Integral term based on output of Sumvox / GainI sub-O ₂ sensor
Sumvox output deviation integrated value
Kgain Integrated gain
GainPr Rich-time proportional term gain
GainPl Lean time proportional term gain

Claims (3)

A catalyst disposed in an exhaust passage of the internal combustion engine;
An upstream exhaust gas sensor disposed upstream of the catalyst;
A downstream side exhaust gas sensor disposed downstream of the catalyst;
Main feedback means for correcting the fuel injection amount based on the output of the upstream side exhaust gas sensor so that the air-fuel ratio upstream of the catalyst becomes the target air-fuel ratio;
Sub-feedback means for correcting the correction by the main feedback means so that the air-fuel ratio downstream of the catalyst becomes the stoichiometric air-fuel ratio based on the output of the downstream exhaust gas sensor,
The sub-feedback means includes
When the downstream exhaust gas sensor emits a rich output, a value obtained by multiplying an output deviation between the output of the downstream exhaust gas sensor and a target value of the output by a rich time proportional term gain is calculated as a proportional term, When the downstream exhaust gas sensor emits a lean output, a value obtained by multiplying the output deviation by a lean time proportional term gain is calculated as a proportional term, and the lean time proportional term gain is calculated as the rich time proportional term. A proportional term calculation means for setting the gain larger than the gain;
Integration term calculation means for calculating an output deviation integrated value by an integration process of integrating a value obtained by multiplying the output deviation by an integration gain every predetermined processing cycle, and calculating an integral term proportional to the output deviation integrated value. ,
The integral term calculation means, compared to lean period cumulative gain used when the downstream exhaust gas sensor is emitting a lean output, the rich period cumulative gain used when the downstream exhaust gas sensor is emitting a rich output And the ratio of the rich-time integrated gain and the lean-time integrated gain is set so that the downstream exhaust gas sensor output is maintained rich and the downstream exhaust gas sensor output is lean. An air-fuel ratio control apparatus for an internal combustion engine, characterized in that the ratio is set to a reciprocal relationship with respect to a ratio to a maintained lean maintenance time . A catalyst disposed in an exhaust passage of the internal combustion engine;
An upstream exhaust gas sensor disposed upstream of the catalyst;
A downstream side exhaust gas sensor disposed downstream of the catalyst;
Main feedback means for correcting the fuel injection amount based on the output of the upstream side exhaust gas sensor so that the air-fuel ratio upstream of the catalyst becomes the target air-fuel ratio;
Sub-feedback means for correcting the correction by the main feedback means so that the air-fuel ratio downstream of the catalyst becomes the stoichiometric air-fuel ratio based on the output of the downstream exhaust gas sensor,
The sub-feedback means includes
When the downstream exhaust gas sensor emits a rich output, a value obtained by multiplying an output deviation between the output of the downstream exhaust gas sensor and a target value of the output by a rich time proportional term gain is calculated as a proportional term, When the downstream exhaust gas sensor emits a lean output, a value obtained by multiplying the output deviation by a lean time proportional term gain is calculated as a proportional term, and the rich time proportional term gain is calculated as the lean time proportional term. A proportional term calculation means for setting the gain larger than the gain;
Integration term calculation means for calculating an output deviation integrated value by an integration process of integrating a value obtained by multiplying the output deviation by an integration gain every predetermined processing cycle, and calculating an integral term proportional to the output deviation integrated value. ,
The integral term calculation means, compared rich period cumulative gain used when the downstream exhaust gas sensor is emitting a rich output, the lean period cumulative gain used when the downstream exhaust gas sensor is emitting a lean output And the ratio of the rich-time integrated gain and the lean-time integrated gain is set so that the downstream exhaust gas sensor output is maintained rich and the downstream exhaust gas sensor output is lean. An air-fuel ratio control apparatus for an internal combustion engine, characterized in that the ratio is set to a reciprocal relationship with respect to a ratio to a maintained lean maintenance time . The integral term calculation means, the ratio between the lean period cumulative gain before Symbol rich period cumulative gain, the said lean time proportional term gain to correspond to the ratio between the rich time proportional term gain, wherein the lean period cumulative The air-fuel ratio control apparatus for an internal combustion engine according to claim 1 or 2, wherein a gain and the rich-time integrated gain are set.

Images