JP2009257188A - Air-fuel ratio control device and air-fuel ratio control method for internal combustion engine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an air-fuel ratio control device and an air-fuel ratio control method for an internal combustion engine capable of always securing good robustness and convergency of learning control irrespective of a condition of over response of feedback learning value during learning control. <P>SOLUTION: In renewal of sub F/B learning value SG during sub F/B learning control, a value of sub F/B learning value SG when deviation amount E from an initial value A which is a value at the start of the renewal gets largest is stored as the maximum deviation value B. Variable setting of proportional gain Kp is done according to return amount C to keep proportional gain Kp of a sub F/B correction value VH smaller, as a value of the return amount C which is deviation of a current value of the sub F/B learning value SG from the maximum deviation value B is larger. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention provides a feedback correction value that is updated in accordance with a difference between an actual value of the control amount and a target value, and a feedback correction value that is manipulated to change a control amount related to the air-fuel ratio of the internal combustion engine. The present invention relates to an air-fuel ratio control apparatus and an air-fuel ratio control method for an internal combustion engine that perform feedback control of the air-fuel ratio by correcting the value with a feedback learning value updated to approach “0”.

  As is well known, a three-way catalytic converter that supports a three-way catalyst using a noble metal such as platinum, palladium, or rhodium as an active material is used as an exhaust gas purification device for an internal combustion engine such as a vehicle. The three-way catalyst simultaneously performs oxidation of carbon monoxide (CO) and hydrocarbon (HC) in the exhaust gas and reduction of nitrogen oxide (NOx), which are harmless carbon dioxide (CO2) and water (H2O). ) And nitrogen (N2) to purify the exhaust gas. Exhaust gas purification using such a three-way catalyst is most effectively performed when the oxygen concentration in the catalyst atmosphere is the stoichiometric air-fuel ratio of the air-fuel ratio burned in the internal combustion engine. Therefore, in an internal combustion engine that employs a three-way catalytic converter as an exhaust purification device, air-fuel ratio feedback control is performed to maintain the air-fuel ratio of the air-fuel mixture to be burned within a certain narrow width (window) near the stoichiometric air-fuel ratio. I am doing so. In such air-fuel ratio feedback control, the fuel injection amount is feedback-corrected based on the output of the oxygen concentration sensor that detects the exhaust oxygen concentration, so that the air-fuel ratio is maintained in the window and a good reaction for exhaust purification is achieved. Creating conditions. In recent years, as seen in, for example, Patent Document 1, empty feedback is performed through double feedback control of main feedback control based on exhaust oxygen concentration on the upstream side of the catalyst and sub feedback control based on exhaust oxygen concentration on the downstream side of the catalyst. Devices for controlling the fuel ratio are also in practical use.

  In such air-fuel ratio feedback control, learning control of feedback correction values is performed. This learning control is performed by storing the value of the feedback correction value when the air-fuel ratio becomes the target value. In the subsequent air-fuel ratio feedback control, the stored value (feedback learning value) is added in advance to achieve early convergence of the feedback.

Note that learning of such feedback learning values needs to be completed during a limited period. Therefore, in the air-fuel ratio control device described in Patent Literature 2, the feedback learning value is converged earlier by increasing the update gain of the feedback learning value as the difference between the feedback correction value and the feedback learning value is larger. ing. Further, in the air-fuel ratio control device described in Patent Document 3, the deviation between the detected value of the oxygen concentration on the downstream side of the catalyst and the target value is integrated, and the update rate of the feedback learning value is made variable according to the integrated value. Thus, the feedback learning value is converged early while suppressing the overshoot of the air-fuel ratio.
Japanese Patent Laid-Open No. 2002-227689 Japanese Patent Laid-Open No. 09-112310 JP 2007-092688 A

  By the way, in learning control as described above, it is necessary to complete learning within a limited period. Therefore, the update rate of the feedback learning value during learning is set to be high to some extent. For this reason, an excessive response (overshoot or undershoot) of the feedback learning value often occurs during the learning process. For example, in the example of the transition of the feedback learning value during learning shown in FIG. 7, after the learning is started, the feedback learning value once undershoots and then gradually converges to an appropriate value α.

  The feedback gain of the feedback correction value greatly affects the convergence of the feedback learning value after such overresponse. For example, in order to achieve early convergence of the feedback learning value after overresponse, it is necessary to set a large feedback gain. However, if the feedback gain is too large, the feedback learning value will oscillate as shown by the curve La in the figure, resulting in unstable control. On the other hand, if the feedback gain is too small, it takes too much time for the feedback learning value to converge as shown by the curve Lb in FIG. Therefore, it is necessary to optimize the feedback gain of the feedback correction value in order to ensure good convergence and stability of the feedback learning value after overresponse. However, the appropriate feedback gain value changes depending on the situation of the excessive response of the feedback learning value at that time. Therefore, in order to set the optimal feedback gain, it is necessary to predict the overresponse characteristics of the feedback learning value, and the convergence and stability of the feedback learning value after overresponse are ensured due to the difficulty of the prediction. It is difficult to do.

  The present invention has been made in view of such circumstances, and the problem to be solved is that the convergence and robustness of learning control are always maintained regardless of the over-response status of the feedback learning value during learning control. An object of the present invention is to provide an air-fuel ratio control device and an air-fuel ratio control method for an internal combustion engine that can be ensured satisfactorily.

Hereinafter, means for solving the above-described problems and the effects thereof will be described.
In order to solve the above-described problem, the invention according to claim 1 is directed to the operation amount operated to change the control amount related to the air-fuel ratio of the internal combustion engine according to the difference between the actual value of the control amount and the target value. In the air-fuel ratio control apparatus for an internal combustion engine that performs feedback control of the air-fuel ratio by correcting the feedback correction value by the feedback correction value updated to be close to “0”, Feedback gain setting for variably setting the feedback gain according to the degree of return so that the feedback gain of the feedback correction value decreases as the return from the overresponse of the feedback learned value progresses when updating the feedback learned value The gist is to provide means.

  In the above configuration, when the feedback learning value over-responds, that is, overshoot or undershoot during learning control, the feedback gain of the subsequent feedback correction value is set smaller as the feedback learning value recovers from the over-response. Become so. Here, the update speed of the feedback learning value updated to bring the feedback correction value closer to “0” increases as the feedback gain of the feedback correction value increases. Therefore, in the above configuration, immediately after the overresponse, the update rate of the feedback learning value is increased, and as the return of the feedback learning value after the overresponse progresses, the update rate becomes slower, and after the overresponse, It is possible to achieve both responsiveness and convergence of the feedback learning value.

  In the above configuration, since the feedback gain is variably set based on the degree of return from overresponse, it is possible to set the feedback gain corresponding to the overresponse characteristics of the feedback learning value at that time. That is, the feedback gain can always be appropriately set even if the feedback response value over-response situation is different. Therefore, according to the above configuration, the convergence and robustness of the learning control can always be ensured satisfactorily regardless of the state of the excessive response of the feedback learning value during the learning control.

  In order to solve the above-mentioned problem, in the invention according to claim 2, the operation amount operated to change the control amount related to the air-fuel ratio of the internal combustion engine is determined according to the difference between the actual value of the control amount and the target value. In the air-fuel ratio control apparatus for an internal combustion engine that performs feedback control of the air-fuel ratio by correcting the feedback correction value by the feedback correction value updated to be close to “0”, When updating the feedback learning value, the storage means for storing the value of the feedback learning value when the deviation from the value at the start of the update becomes maximum, the value stored by the storage means, and the feedback learning value The feedback gain of the feedback correction value decreases as the deviation from the current value increases. A feedback gain setting means for the-back gain is variably set, in that it comprises has as its gist.

  In the above configuration, the value of the feedback learning value when the amount of deviation from the value at the start of the update becomes the maximum, that is, the value at the time of overresponse of the feedback learning value is stored. After that, the feedback gain of the feedback correction value is set to be smaller as the deviation between the stored value and the current value of the feedback learning value increases, that is, as the feedback learning value converges after overresponse. Become. Therefore, in the above configuration, immediately after the overresponse, the update rate of the feedback learning value is increased, and as the feedback learned value converges after the overresponse progresses, the update rate becomes slower, and after the overresponse, It is possible to achieve both responsiveness and convergence of the feedback learning value.

  In the above configuration, since the feedback gain is variably set based on the value of the feedback learning value at the time of overresponse, it is possible to set the feedback gain corresponding to the overresponse characteristics of the feedback learning value at that time. That is, the feedback gain can always be appropriately set even if the feedback response value over-response situation is different. Therefore, according to the above configuration, the convergence and robustness of the learning control can always be ensured satisfactorily regardless of the state of the excessive response of the feedback learning value during the learning control.

  According to a third aspect of the present invention, in the air-fuel ratio control apparatus for an internal combustion engine according to the first or second aspect, the air-fuel ratio control apparatus grasps the output from the air-fuel ratio sensor for detecting the exhaust oxygen concentration upstream of the catalyst. Main feedback control for feedback adjustment of the fuel injection amount so as to reduce the difference between the target air-fuel ratio and the target air-fuel ratio, and the air-fuel ratio to reduce the deviation between the measured value of the exhaust oxygen concentration downstream of the catalyst and the target value. The feedback control of the air-fuel ratio is performed through sub-feedback control for feedback adjustment of the output of the fuel ratio sensor, the operation amount is the exhaust oxygen concentration on the downstream side of the catalyst, and the operation amount is the output of the air-fuel ratio sensor. The gist is that it is.

  In the above-described configuration, the learning during the learning control in the air-fuel ratio control apparatus for an internal combustion engine according to each of the above claims is targeted for the sub-feedback learning control in the case of performing the above-described main / sub dual feedback control. The feedback gain is variably set according to the response status.

  In order to solve the above-mentioned problem, in the invention according to claim 4 as an air-fuel ratio control method for an internal combustion engine, an operation amount operated to change a control amount related to the air-fuel ratio of the internal combustion engine is set to an actual value of the control amount. A method of performing feedback control of the air-fuel ratio by correcting with a feedback correction value updated according to a difference between a value and a target value and a feedback learning value updated to bring the feedback correction value close to “0” When the feedback learning value is updated, the feedback gain is variable according to the degree of the return so that the feedback gain of the feedback correction value becomes smaller as the return from the excessive response of the feedback learning value proceeds. The gist is that it was set.

  In the above method, when the feedback learning value is over-responsive during learning control, that is, overshoot or undershoot, the feedback gain of the subsequent feedback correction value is set smaller as the feedback learning value recovers from the over-response. It becomes like this. Here, the update speed of the feedback learning value updated to bring the feedback correction value closer to “0” increases as the feedback gain of the feedback correction value increases. Therefore, in the above method, the update rate of the feedback learning value is increased immediately after the overresponse, and the update rate becomes slower as the return of the feedback learned value after the overresponse progresses. It is possible to achieve both responsiveness and convergence of the feedback learning value.

  In the above method, since the feedback gain is variably set based on the over-response time, it is possible to set the feedback gain corresponding to the over-response characteristic of the feedback learning value at that time. That is, the feedback gain can always be appropriately set even if the feedback response value over-response situation is different. Therefore, according to the above method, the convergence and robustness of the learning control can always be ensured satisfactorily regardless of the state of the excessive response of the feedback learning value during the learning control.

  In order to solve the above-mentioned problem, an invention according to claim 5 as an air-fuel ratio control method for an internal combustion engine is characterized in that an operation amount operated to change a control amount related to the air-fuel ratio of the internal combustion engine A method of performing feedback control of the air-fuel ratio by correcting with a feedback correction value updated according to a difference between a value and a target value and a feedback learning value updated to bring the feedback correction value close to “0” When the feedback learning value is updated, the feedback learning value when the deviation from the value at the start of updating is maximized is stored, and the stored value and the feedback learning value are The feedback according to the deviation is reduced so that the feedback gain of the feedback correction value decreases as the deviation from the current value increases. In that the gain is variably set has its gist.

  In the above method, the value of the feedback learning value when the amount of deviation from the value at the start of the update becomes the maximum, that is, the value at the time of an overresponse of the feedback learning value is stored. After that, the feedback gain of the feedback correction value is set to be smaller as the deviation between the stored value and the current value of the feedback learning value increases, that is, as the return of the feedback learning value after overresponse progresses. Become. Therefore, immediately after the overresponse, the update rate of the feedback learning value is increased, and as the feedback learning value converges after the overresponse progresses, the update rate becomes slower, and the feedback learning value after the overresponse becomes lower. Both responsiveness and convergence can be achieved.

  In the above method, since the feedback gain is variably set based on the value of the feedback learning value at the time of overresponse, it is possible to set the feedback gain corresponding to the overresponse characteristics of the feedback learning value at that time. That is, the feedback gain can always be appropriately set even if the feedback response value over-response situation is different. Therefore, according to the above method, the convergence and robustness of the learning control can always be ensured satisfactorily regardless of the state of the excessive response of the feedback learning value during the learning control.

  According to a sixth aspect of the present invention, in the air-fuel ratio control method for an internal combustion engine according to the fourth or fifth aspect, the air-fuel ratio feedback control is based on an output of an air-fuel ratio sensor that detects the exhaust oxygen concentration upstream of the catalyst. Main feedback control for feedback adjustment of the fuel injection amount in order to reduce the difference between the grasped air-fuel ratio and the target air-fuel ratio, and the above-mentioned in order to reduce the deviation between the measured value of the exhaust oxygen concentration downstream of the catalyst and the target value. It is performed through sub-feedback control for feedback adjustment of the output of the air-fuel ratio sensor, the manipulated variable is the exhaust oxygen concentration on the downstream side of the catalyst, and the manipulated variable is the output of the air-fuel ratio sensor. It is said.

  The method described above is intended for sub-feedback learning control when performing the above-described main / sub dual feedback control. The feedback gain is variably set according to the response status.

  DESCRIPTION OF EMBODIMENTS Hereinafter, an embodiment embodying an air-fuel ratio control apparatus and an air-fuel ratio control method for an internal combustion engine according to the present invention will be described in detail with reference to FIGS. Note that the air-fuel ratio control device and the air-fuel ratio control method of the present embodiment are applied to an in-vehicle internal combustion engine.

  In the internal combustion engine shown in FIG. 1, the amount of air taken into the combustion chamber 12 is adjusted through opening control of the throttle valve 11 installed in the intake passage 10. An air-fuel mixture of this air and fuel injected from the injector 13 is combusted in the combustion chamber 12. Exhaust gas generated by the combustion of the air-fuel mixture in the combustion chamber 12 is discharged to the exhaust passage 14 and is exhausted by the exhaust purification catalyst of the two catalytic converters (front catalytic converter 15 and rear catalytic converter 16) installed in the exhaust passage 14. Purified. In these catalytic converters (15, 16), as an exhaust purification catalyst, there is a three-way catalyst for purifying these by simultaneously oxidizing HC and CO and reducing NOx, and a promoter such as cerium having oxygen storage capacity. A higher exhaust purification performance is exhibited by the oxygen storage capacity of the promoter.

  Such air-fuel ratio feedback control is executed by the electronic control unit 17 that controls the engine. The electronic control unit 17 temporarily stores a central processing unit (CPU) that executes various arithmetic processes related to engine control, a read-only memory (ROM) in which engine control programs and data are stored, CPU arithmetic results, and the like. And a random access memory (RAM) stored in the memory, an input / output port (I / O) for inputting / outputting signals to / from the outside, and the like.

  The input port of the electronic control unit 17 includes an accelerator sensor 18 that detects the accelerator operation amount, a throttle sensor 19 that detects the opening degree of the throttle valve 11, an air flow meter 20 that detects the intake air amount, and an engine speed. The output of the fuel pressure sensor 22 that detects the pressure (fuel pressure) of the fuel supplied to the NE sensor 21 and the injector 13 is input.

  In addition, an air-fuel ratio sensor 23 installed on the upstream side of the catalyst (exhaust upstream side of the front catalytic converter 15) and an downstream side of the catalyst (exhaust downstream side of the rear catalytic converter 16) are installed at the input port of the electronic control unit 17. The output of the oxygen sensor 24 is also input. As shown in FIG. 2, the air-fuel ratio sensor 23 outputs a voltage signal corresponding to the exhaust oxygen concentration upstream of the catalyst. The air-fuel ratio sensor 23 employed in the present embodiment has an output VAF of just “0 V” when the exhaust gas oxygen concentration on the upstream side of the catalyst is the value X when the air-fuel mixture is combusted at the stoichiometric air-fuel ratio. It is comprised so that. The oxygen sensor 24 outputs a voltage signal corresponding to the exhaust oxygen concentration on the downstream side of the catalyst. As shown in FIG. 3, the output VO of the oxygen sensor 24 changes in a step-like manner with the boundary when the exhaust oxygen concentration downstream of the catalyst is the value X when the air-fuel mixture is burned at the stoichiometric air-fuel ratio. It is supposed to be. The oxygen sensor 24 employed in the present embodiment is configured such that its output VO becomes “0.5 V” when combustion is performed at the stoichiometric air-fuel ratio.

  Hereinafter, details of the air-fuel ratio feedback control executed by the electronic control unit 17 will be described. Here, double feedback control of main feedback control and sub feedback control is performed as air-fuel ratio feedback control. In the main feedback, feedback adjustment of the fuel injection amount is performed so as to reduce the difference between the air-fuel ratio obtained from the output VAF of the air-fuel ratio sensor 23 that detects the exhaust gas oxygen concentration upstream of the catalyst and the target air-fuel ratio. It has become. In the sub-feedback control, feedback adjustment of the output VAF of the air-fuel ratio sensor 23 is performed so as to reduce the deviation between the measured value of the exhaust oxygen concentration downstream of the catalyst by the oxygen sensor 24 and the target value.

  The electronic control unit 17 basically calculates the fuel injection amount required at that time as the command injection amount Q based on the engine rotation speed and the engine load factor, and calculates the calculated command injection amount Q. The injector 13 is driven so that fuel is injected. The engine speed used for calculating the command injection amount Q at this time is obtained from the detection result of the NE sensor 21. The engine load factor indicates the ratio of the current load to the maximum load of the internal combustion engine, and is calculated based on the engine speed and the intake air amount detected by the air flow meter 20.

  When the injector 13 is driven in accordance with the command injection amount Q, the command injection time TAU is calculated as the energization time of the injector 13 necessary for injecting fuel for the command fuel injection amount Q. The command injection time TAU is calculated using the following equation (1).

TAU = Q × K1 × KINJA + KINJB (1)

In the above equation (1), “K1” is the change in the fuel injection amount due to the difference in the fuel injection rate (the fuel injection amount from the injector 13 per unit time) corresponding to the pressure (fuel pressure) of the fuel supplied to the injector 13. Is a fuel pressure correction coefficient for compensating The value of the fuel pressure correction coefficient K1 is set to “1.0” when the fuel pressure is a specified reference fuel pressure, and the value increases as the fuel pressure decreases, and the value decreases as the fuel pressure increases. It has become. In addition, “KINJA” in the above equation (1) is a sensitivity coefficient, and the value indicates the energization time of the injector 13 necessary for unit amount fuel injection when the fuel pressure is the reference fuel pressure. Further, “KINJB” in the above formula (1) is an invalid injection period of the injector 13, and its value corresponds to the time from the start of energization to the injector 13 until the actual fuel injection is started. Yes.

  Next, the details of the procedure for calculating the command injection amount Q will be described. The command injection amount Q includes a basic injection amount Qbase, a main feedback correction value (hereinafter referred to as “main F / B correction value DF”), and a main feedback learning value (hereinafter referred to as “main F / B learning value MG (i)”). Based on the description (2).

Q = Qbase + DF + MG (i) (2)

Here, the basic injection amount Qbase is a theoretical value of the fuel injection amount necessary for setting the air-fuel ratio of the air-fuel mixture burned in the combustion chamber 12 to the stoichiometric air-fuel ratio (= 14.7). It is obtained as a value (GA / 14.7) obtained by dividing the intake air amount GA detected by the air flow meter 20 by the theoretical air-fuel ratio. The main F / B correction value DF is a correction value that is increased or decreased in accordance with the deviation between the actual air-fuel ratio and the stoichiometric air-fuel ratio that is grasped from the detection result of the exhaust gas oxygen concentration upstream of the catalyst by the air-fuel ratio sensor 23. The command injection amount Q and consequently the command injection time TAU are increased or decreased so that the actual air fuel ratio is maintained in the vicinity of the theoretical air fuel ratio by increasing or decreasing the main F / B correction value DF. Further, the main F / B learning value MG (i) is a steady state between the theoretical value and the actual value of the fuel injection amount required to obtain the theoretical air-fuel ratio due to individual differences in the fuel system and intake system of the internal combustion engine and deterioration over time. The deviation is compensated, and the value is updated according to the value of the main F / B correction value DF.

  More specifically, the main F / B correction value DF is calculated using the following equation (3) based on the fuel amount deviation ΔQ, the proportional gain Gp, the fuel amount deviation integrated value ΣΔQ, and the integral gain Gi. The

DF = ΔQ × Gp + ΣΔQ × Gi (3)

The first term “ΔQ × Gp” on the right side of Equation (3) is a proportional term that takes a value proportional to the amount of deviation of the actual air-fuel ratio from the theoretical air-fuel ratio. The fuel amount deviation ΔQ in the proportional term is calculated using the following equation (4) based on the intake air amount GA, the actual air-fuel ratio ABF, and the basic injection amount Qbase. The proportional gain Gp is a constant whose value is adapted beforehand through experiments and simulations, and is set to a negative value here.

ΔQ = (GA / ABF) −Qbase (4)

Note that the value of the actual air-fuel ratio ABF in the above equation (4) is calculated from the output VAF of the air-fuel ratio sensor 23 (strictly speaking, the output VAFs after correction in the sub feedback control). The value of the fuel amount deviation ΔQ thus obtained is a value obtained by subtracting the theoretical fuel amount necessary for obtaining the theoretical air-fuel ratio from the actually burned fuel amount.

  Further, the second term “ΣΔQ × Gi” on the right side of the above equation (3) is an integral for eliminating the residual deviation of the actual air-fuel ratio from the stoichiometric air-fuel ratio, which cannot be solved only by the proportional term “ΔQ × Gp”. It is a term. The fuel amount deviation integral value ΣΔQ here is obtained as a time integral value of the fuel amount deviation ΔQ. The integral gain Gi here is a constant whose value is adapted beforehand through experiments, simulations, and the like, and is set to a negative value here.

  The main F / B correction value DF thus obtained basically decreases the basic injection amount Qbase when combustion is performed at an air-fuel ratio richer than the stoichiometric air-fuel ratio and the exhaust oxygen concentration on the upstream side of the catalyst becomes low. When the combustion is performed at an air-fuel ratio leaner than the stoichiometric air-fuel ratio and the exhaust oxygen concentration on the upstream side of the catalyst becomes high, the value is increased or decreased so as to increase the basic injection amount Qbase. Therefore, the air-fuel ratio of the air-fuel mixture burned in the combustion chamber 12 is feedback-adjusted so as to approach the stoichiometric air-fuel ratio through such increase / decrease of the main F / B correction value DF.

  On the other hand, the main F / B learning value MG (i) described above is updated in accordance with the main F / B correction value DF calculated as described above. Specifically, the main F / B learning value MG (i) is updated to the main F / B correction value DF at that time when both the following update conditions (A) and (B) are satisfied. It has come to be.

  (A) The correction rate of the basic injection amount Qbase by the main F / B correction value DF, that is, the ratio of the main F / B correction value DF to the basic injection amount Qbase (= | DF / Qbase |) is sufficiently large (for example, “1 %"more than).

(B) The fluctuation amount of the main F / B correction value DF is sufficiently small.
The main F / B correction value DF comes closer to “0” depending on the result of correction using the main F / B learning value MG (i). The value of the main F / B learning value MG (i) when the main F / B correction value DF approaches “0” corresponds to the stoichiometric air-fuel ratio due to individual differences in the intake system and the fuel system and changes over time. The value corresponds to the steady deviation of the actual air-fuel ratio.

  The main F / B learning value MG (i) is calculated individually for each of a plurality of learning regions i (i = 1, 2, 3,...) Divided according to the load region of the internal combustion engine. Then, according to the engine load at that time, the corresponding learning region i is selected as the main F / B learning value MG (i) to be used.

  As described above, in the fuel injection control device for an internal combustion engine according to the present embodiment, the air-fuel ratio feedback control based on the detection result of the air-fuel ratio sensor 23 that detects the exhaust oxygen concentration upstream of the catalyst, that is, the main feedback control is performed. It is done. On the other hand, in the present embodiment, in addition to this, sub-feedback control is performed in order to suppress a decrease in accuracy of the main feedback control due to variations in output characteristics of the air-fuel ratio sensor 23 due to individual differences and changes over time.

  In such sub-feedback control, the output VAF of the air-fuel ratio sensor 23 used to calculate the actual air-fuel ratio ABF used for the main feedback control is corrected in the manner shown in the following equation (5). In the following equation (5), “VAFs” indicates the output of the air-fuel ratio sensor 23 after the correction by the sub feedback control, and “VAF” indicates the output before the correction.

VAFs → VAF + VH + SG (5)

“VH” in the above equation (5) is a sub feedback (sub F / B) correction value, and the value is increased or decreased according to the detection result of the oxygen sensor 24 installed downstream of the catalyst. Yes. Further, “SG” in the above equation (5) is a sub F / B learning value, and the value is updated according to the increase / decrease of the sub F / B correction value VH.

  The sub F / B correction value VH is calculated using the following equation (6) based on the voltage deviation ΔV, the proportional gain Kp, the voltage deviation integral value ΣΔV, the integral gain Ki, the voltage differential value dV, and the differential gain Kd.

VH = ΔV × Kp + ΣΔV × Ki + dV × Kd (6)

The first term “ΔV × Kp” on the right side of equation (6) is proportional to the deviation between the actual value of the exhaust oxygen concentration on the downstream side of the catalyst and the value when combustion is performed at the stoichiometric air-fuel ratio. It is a proportional term that takes a value. The voltage deviation ΔV in the proportional term “ΔV × Kp” is calculated as a value (= VO−tVO) obtained by subtracting the target output tVO from the actual output VO of the oxygen sensor 24. Here, the target output tVO is set to the theoretical value of the output of the oxygen sensor 24 when combustion is performed at the stoichiometric air-fuel ratio. As shown in FIG. 3, when combustion is performed at an air-fuel ratio leaner than the stoichiometric air-fuel ratio and the exhaust oxygen concentration on the downstream side of the catalyst becomes higher than the value X at the stoichiometric air-fuel ratio, the oxygen sensor 24 The output VO is smaller than the target output tVO. When combustion is performed at an air-fuel ratio richer than the stoichiometric air-fuel ratio and the exhaust gas oxygen concentration on the downstream side of the catalyst becomes lower than the value X at the stoichiometric air-fuel ratio, the output VO takes a value larger than the target output tVO. It has become. Further, the proportional gain Kp in the proportional term “ΔV × Kp” is a constant whose value is adapted in advance through experiments and simulations, and is set to a negative value here.

  Further, the second term “ΣΔV × Ki” on the right side of the above equation (6) is an integral for eliminating the residual deviation of the actual air-fuel ratio with respect to the theoretical air-fuel ratio, which cannot be solved only by the proportional term “ΔV × Kp”. It is a term. The voltage deviation integral value ΣΔV here is obtained as a time integral value of the voltage deviation ΔV. Further, the integral gain Ki here is a constant whose value is adapted in advance through experiments and simulations, and is set to a negative value here.

  Further, the third term “dV × Kd” on the right side of the above equation (6) is a differential term for enhancing the response when the exhaust oxygen concentration on the downstream side of the catalyst is converged to the value at the stoichiometric air-fuel ratio. The voltage differential value dV here is obtained as a time differential value of the output VO of the oxygen sensor 24, and the value represents the amount of change per unit time of the output VO. The differential gain Kd in the differential term “dV × Kd” is a constant whose value is adapted beforehand through experiments, simulations, and the like, and is set to a negative value here.

  The sub F / B correction value VH obtained in this way is the value of the output VO of the oxygen sensor 24 when the combustion is performed at an air-fuel ratio richer than the stoichiometric air-fuel ratio and the exhaust oxygen concentration on the downstream side of the catalyst becomes low. Since it becomes larger than the output tVO, its value is decreased so as to decrease and correct the output VAF of the air-fuel ratio sensor 23 used for calculation of the actual air-fuel ratio ABF. Therefore, the calculated value of the actual air-fuel ratio ABF at this time is corrected to a value indicating a richer air-fuel ratio, and the main F / B correction value DF is so corrected that the basic injection amount Qbase is further reduced. The value of will be decreased. On the other hand, when combustion is performed at an air-fuel ratio leaner than the stoichiometric air-fuel ratio and the exhaust gas oxygen concentration on the downstream side of the catalyst becomes high, the output VO of the oxygen sensor 24 becomes smaller than the target output tVO, so that the actual air-fuel ratio ABF The value is increased so as to increase and correct the output VAF of the air-fuel ratio sensor 23 used for the calculation of the above. Therefore, the calculated value of the actual air-fuel ratio ABF at this time is corrected to a value indicating a leaner air-fuel ratio, and the main F / B correction value DF so that the basic injection amount Qbase is further corrected to increase. The value of is increased. Therefore, depending on the increase / decrease in the sub F / B correction value VH, the fuel injection amount is feedback adjusted so that the exhaust oxygen concentration on the downstream side of the catalyst becomes the value when combustion is performed at the stoichiometric air-fuel ratio. Become.

  On the other hand, the update of the sub F / B learning value SG according to the increase / decrease of the sub F / B correction value VH is performed in the following manner. That is, when updating the sub F / B learning value SG, first, a gradual change value of the sub F / B correction value VH is calculated, and further, an upper limit guard and a lower limit guard are applied to the gradual change value to calculate the update amount SGK. The Then, the update amount SGK is set as the latest value obtained by adding the update amount SGK to the sub F / B learning value SG before update, so that the sub F / B learning value SG is updated ("updated SG"). ← “SG before update” + SGK).

  Depending on the sub F / B learning value SG updated in this way, the value of the sub F / B correction value VH becomes closer to “0”. Then, the value of the sub F / B learning value SG when the value of the sub F / B correction value VH approaches “0” depends on individual differences in the output characteristics of the air-fuel ratio sensor 23 and the exhaust gas purification characteristics of the catalyst, and changes over time. The resulting value corresponds to the steady-state deviation of the actual air-fuel ratio with respect to the stoichiometric air-fuel ratio. Hereinafter, such control related to the update of the sub F / B learning value SG is referred to as “sub F / B learning control”.

  Next, details of sub F / B learning control in this embodiment will be described. In the present embodiment, when updating the sub F / B learning value SG during sub F / B learning control, the value of the sub F / B learning value SG at the time of overresponse, that is, at the time of overshoot or undershoot is used as a reference. The feedback gain of the sub F / B correction value VH (hereinafter referred to as “F / B gain”) is variably set. As a result, the convergence and robustness of the learning control are always ensured satisfactorily regardless of the over-response state of the sub F / B learning value SG during the sub F / B learning control. Specifically, when the sub F / B learning value SG is updated, the value of the sub F / B learning value SG when the amount of deviation from the value at the start of the updating becomes maximum is stored. . Then, as the deviation between the stored value and the current value of the sub F / B learning value SG is larger, the F / B gain of the sub F / B correction value VH becomes smaller, so that the F / B according to the deviation. The gain is variably set. As a result, when the sub F / B learning value SG is updated, the F / B gain of the sub F / B correction value VH is set to be smaller as the return from the excessive response of the sub F / B learning value SG proceeds. I am doing so.

  In this embodiment, the variable setting of the F / B gain of the sub F / B correction value VH during the sub F / B learning control is performed for the proportional gain Kp. Of course, if necessary, the integral gain Ki and the differential gain Kd may also be subject to such variable setting.

  FIG. 4 shows an example of the control mode of the sub F / B learning control in this embodiment. In the example of the figure, the sub F / B learning value SG is decreased after the sub F / B learning control is started. Then, the sub F / B learning value SG once increases under the appropriate value α and then increases so that it gradually converges to the appropriate value α. Here, in the present embodiment, when the amount of deviation E (= | SGF−SG |) from the value at the start of updating the sub F / B learning value SG (initial value A) becomes the maximum, In the example, the value of the sub F / B learning value SG when the sub F / B learning value SG becomes the minimum is stored as the maximum deviation value B. Then, after the return from the undershoot starts and the sub F / B learning value SG starts to increase, the return amount C (which is the deviation between the maximum deviation value B and the current value of the sub F / B learning value SG) = | B−SG |), the proportional gain Kp of the sub F / B correction value VH is variably set.

  The proportional gain Kp is set by calculating the proportional gain Kp based on the return amount C using a one-dimensional map of the return amount C and the proportional gain Kp as illustrated in FIG. This one-dimensional map is created so that the value of the proportional gain Kp increases as the value of the return amount C decreases. Therefore, immediately after the start of the recovery of the sub F / B learning value SG from the undershoot, the value of the proportional gain Kp is set to be large, and thereafter, as the recovery from the undershoot proceeds, the value of the proportional gain Kp is gradually decreased. Will be set. In general, as the value of the proportional gain Kp increases, the response speed of the sub F / B correction value VH increases and the update speed of the sub F / B learning value SG tends to increase. Therefore, immediately after the undershoot, the update speed of the sub F / B learning value SG is increased, and the update speed becomes slower as the recovery from the over-response progresses. Therefore, it is possible to achieve both responsiveness and convergence of the sub F / B learning value SG after undershoot. Even when the sub F / B learning value SG overshoots after the start of updating, the sub F / B learning value SG is basically changed in the opposite direction, and the sub F / B learning value SG is basically the same as undershooting. / B learning control is performed.

  Here, in the present embodiment, the value of the proportional gain Kp after the over response (undershoot / overshoot) of the sub F / B learning value SG is used as a reference based on the value of the sub F / B learning value SG at the time of over response. Variable setting. Therefore, the proportional gain Kp can always be set appropriately even if the excessive response status of the sub F / B learning value SG is different. That is, the proportional gain Kp can be appropriately set corresponding to the over-response characteristics at that time. For this reason, in the present embodiment, the convergence and robustness of learning control can always be ensured satisfactorily regardless of the over-response situation of the sub F / B learning value SG during sub F / B learning. It has become.

  FIG. 6 shows a flowchart of an F / B gain setting routine related to variable setting of the proportional gain Kp during such sub F / B learning control. The processing of this routine is repeatedly executed by the electronic control unit 17 every prescribed control period during the execution of the sub F / B learning control.

  When the processing of this routine is started, the electronic control unit 17 first checks in step S10 whether or not the sub F / B learning execution flag has just been switched from “0” to “1”. . The sub F / B learning execution flag is set to “1” when the sub F / B learning control is started, and is cleared to “0” when the sub F / B learning control is completed. Therefore, the electronic control unit 17 at this time confirms whether or not it is immediately after the start of the sub F / B learning control.

  Here, if the electronic control unit 17 is immediately after the start of the sub F / B learning control (S10: YES), after setting the value of the sub F / B learning value SG at that time to the initial value A in step S20. Then, the process proceeds to step S30. On the other hand, if it is not immediately after the start of the sub F / B learning control (S10: NO), the electronic control unit 17 skips the process of step S20 and advances the process to step S30 as it is.

  When the process proceeds to step S30, the electronic control unit 17 determines that the deviation amount E (= | A−SG |) of the current sub F / B learning value SG from the initial value A is the deviation up to that point in step S30. It is confirmed whether or not it is equal to or greater than the maximum value of the amount E (maximum deviation amount Emx). Here, while the sub F / B learning value SG continues to increase / decrease from the start of the update, the deviation amount E gradually increases. When the sub F / B learning value SG starts to recover from overshoot / undershoot, the deviation amount E starts to decrease. Therefore, if the divergence amount E is equal to or greater than the maximum divergence amount Emx, the sub F / B learning value SG is in a period before the start of recovery from overresponse, and the divergence amount E is less than the maximum divergence amount Emx. If it becomes, it will be in the period after the sub F / B learning value SG starts the return | restoration from an overresponse.

  If the deviation amount E at that time is equal to or larger than the maximum deviation amount Emx so far (S30: YES), the electronic control unit 17 advances the processing to step S40. In step S40, the electronic control unit 17 updates the value of the maximum deviation amount Emx to the value of the deviation amount E at that time, and further in step S50, changes the maximum deviation value B to the sub F / B learning value SG at that time. After updating to this value, the processing of this routine is terminated. Note that the state where the deviation amount E is equal to or greater than the maximum deviation amount Emx as described above continues until immediately before the sub F / B learning value SG starts to recover from the overresponse. Therefore, finally, the value of the maximum divergence value B is when the divergence amount E from the value at the start of the update (initial value A) becomes maximum, that is, when the peak of overshoot or undershoot is reached. The sub F / B learning value SG is stored.

  On the other hand, if the deviation amount E at that time is less than the maximum deviation amount Emx so far (S30: NO), the electronic control unit 17 advances the process to step S60. In step S60, the electronic control unit 17 determines the return amount C, that is, the deviation (= | B−SG |) between the stored maximum deviation value B and the current value of the sub F / B learning value SG. After setting the proportional gain Kp using the above-described one-dimensional map (FIG. 5), the process of this routine is terminated. Therefore, after the sub F / B learning value SG starts to return from the overresponse, the sub F / B correction is performed according to the return amount C, that is, the degree of return from the over response of the sub F / B learning value SG. The proportional gain Kp of the value VH is variably set. The variable setting of the proportional gain Kp is continued until it is determined that the sub F / B correction value VH has sufficiently approached “0” and learning of the sub F / B learning value SG has been completed.

  In the present embodiment, the output VAF of the air-fuel ratio sensor 23 is equal to the above-mentioned “control amount relating to the air-fuel ratio of the internal combustion engine”, the detected value of the exhaust oxygen concentration downstream of the catalyst by the oxygen sensor 24, more strictly, the same oxygen The output VO of the sensor 24 corresponds to the “operation amount operated to change the control amount”. Further, the sub F / B correction value VH is changed to the “feedback correction value updated in accordance with the difference between the actual value of the control amount and the target value”, and the sub F / B learning value SG is changed to the “feedback correction value“ 0 ”. "Feedback learning value updated to be close to" ". Further, the process of the electronic control unit 17 in step S50 of the F / B gain setting routine is the process performed by the “storage means”, and the process of the electronic control unit 17 in step S60 of the routine is the process of the “feedback gain setting means”. Each process corresponds to a process to be performed.

According to the air-fuel ratio control apparatus and air-fuel ratio control method for an internal combustion engine according to the present embodiment described above, the following effects can be achieved.
In the present embodiment, when the electronic control unit 17 updates the sub F / B learning value SG, the electronic control unit 17 uses the sub F / B when the deviation amount E from the update start value (initial value A) becomes maximum. The value of the B learning value SG is stored as the maximum deviation value B. The electronic control unit 17 then increases the F / B of the sub F / B correction value VH as the return amount C, which is the deviation between the stored maximum divergence value B and the current value of the sub F / B learning value SG, increases. The proportional gain Kp is variably set according to the deviation so that the proportional gain Kp, which is one of the gains, becomes small. Thus, the proportional gain Kp is variably set according to the degree of the return so that the proportional gain Kp becomes smaller as the return from the excessive response of the sub F / B learning value SG proceeds. In this embodiment, when the sub F / B learning value SG is over-responsive during the sub F / B learning control, that is, overshoots or undershoots, the F / B gain of the subsequent sub F / B correction value VH ( The proportional gain Kp) is set to be smaller as the recovery of the sub F / B learning value SG from the excessive response proceeds. Therefore, in the above configuration, immediately after the overresponse, the update rate of the sub F / B learning value SG is increased, and as the return of the sub F / B learning value SG after the overresponse progresses, the update rate becomes slower. As a result, both the responsiveness and convergence of the sub-F / B learning value SG after overresponse can be achieved. In addition, such a variable setting of the F / B gain (proportional gain Kp) is performed based on the degree of return from the overresponse based on the sub F / B learning value SG at the time of the overresponse. Therefore, the F / B gain can always be set appropriately even if the situation of the overresponse of the sub F / B learning value SG is different, corresponding to the overresponse characteristics of the sub F / B learning value SG at that time. The F / B gain can be set. Therefore, according to the present embodiment, the convergence and robustness of the sub F / B learning control can always be ensured satisfactorily regardless of the excessive response status of the sub F / B learning value SG during the learning control. become able to.

In addition, the said embodiment can also be changed and implemented as follows.
In the above embodiment, the variable setting corresponding to the degree of return from the excessive response of the sub F / B learning value SG during the sub F / B learning control is set only for the proportional gain Kp of the sub F / B correction value VH. However, if necessary, the same variable setting may be performed for the integral gain Ki and the differential gain Kd of the sub F / B correction value VH.

  In the above embodiment, the degree of return from the excessive response of the sub F / B learning value SG is the value of the return amount C that is the deviation between the maximum deviation value B and the current value of the sub F / B learning value SG. Based on the above, the F / B gain of the sub F / B correction value VH is variably set. It is also possible to variably set the F / B gain by confirming the degree of return from the excessive response of the sub F / B learning value SG from parameters other than the deviation (recovery amount C). For example, the ratio of the deviation amount (= | A−SG |) between the initial value A of the sub F / B learning value SG and its current value to the maximum deviation amount Emx is returned from the overresponse of the sub F / B learning value SG. It is also possible to variably set the F / B gain by using it as a parameter for measuring the degree of. In this case, the F / B gain of the sub F / B correction value VH is set smaller as the ratio becomes smaller.

  In the above embodiment, when the sub F / B learning value SG is updated, the F / B gain of the sub F / B correction value VH is set according to the degree of return from the excessive response of the sub F / B learning value SG. It was set to be variable. If it is required to ensure convergence and stability after overresponse of the main F / B learning value MG (i) in the learning control of the main F / B learning value MG (i), The F / B gain can be variably set for the F / B gain (proportional gain Gp or integral gain Gi) of the main F / B correction value DF. In this case, the actual air-fuel ratio obtained from the detection result of the exhaust oxygen concentration upstream of the catalyst by the air-fuel ratio sensor 23 is the above-mentioned “control amount relating to the air-fuel ratio of the internal combustion engine”, and the command injection amount Q is the “control”. This corresponds to the “operation amount operated to change the amount”. In addition, the main F / B correction value DF is changed to the “feedback correction value updated according to the difference between the actual value of the control amount and the target value”, and the main F / B learning value MG (i) is changed to the “feedback correction value”. Corresponds to the “feedback learning value updated so as to be close to“ 0 ””.

  The present invention can also be applied to an air-fuel ratio control apparatus and an air-fuel ratio control method that perform air-fuel ratio feedback control in a mode other than the air-fuel ratio feedback control through main / sub double feedback as in the above embodiment. . In short, any device or method that performs air-fuel ratio feedback control that satisfies the following conditions (A) to (C) regardless of which parameter is used as the control amount or the operation amount of feedback control. The present invention can be applied. As a result, the convergence and robustness of the learning control can always be ensured satisfactorily regardless of the over-response status of the feedback learning value being updated.

(A) The feedback correction value is updated according to the difference between the actual value of the control amount related to the air-fuel ratio of the internal combustion engine and the target value.
(B) The feedback learning value is updated so that the feedback correction value approaches “0”.

  (C) The feedback control of the air-fuel ratio is performed by correcting the manipulated variable that is manipulated to change the controlled variable based on the feedback correction value and the feedback learned value.

  The air-fuel ratio control device and the air-fuel ratio control method of the present invention are the number of sensors and catalytic converters as long as the internal combustion engine performs air-fuel ratio feedback control that satisfies the above conditions (A) to (C). It is also possible to apply the present invention to an internal combustion engine having a different configuration from that of the above embodiment.

BRIEF DESCRIPTION OF THE DRAWINGS Schematic which shows typically the structure of the internal combustion engine used as the application object about one Embodiment of this invention. The graph which shows the output characteristic of the air fuel ratio sensor employ | adopted as the same embodiment. The graph which shows the output characteristic of the oxygen sensor employ | adopted as the same embodiment. The time chart which shows an example of the control mode at the time of sub F / B learning control in the embodiment. The graph which shows the relationship between the return amount in the map about an example of the one-dimensional map for proportional gain calculation employ | adopted by the embodiment, and a proportional gain. The flowchart which shows the process sequence of the F / B gain setting routine applied to the embodiment. The time chart which illustrates transition of the feedback learning value at the time of learning control in the conventional air fuel ratio control device.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Intake passage, 11 ... Throttle valve, 12 ... Combustion chamber, 13 ... Injector, 14 ... Exhaust passage, 15 ... Front catalytic converter, 16 ... Rear catalytic converter, 17 ... Electronic control unit (memory | storage means, feedback gain setting means) , 18 ... accelerator sensor, 19 ... throttle sensor, 20 ... air flow meter, 21 ... NE sensor, 22 ... fuel pressure sensor, VAF ... output of air-fuel ratio sensor (control amount), VO ... output of oxygen sensor (operation amount), VH ... sub F / B correction value (feedback correction value), SG ... sub F / B learning value (feedback learning value).

Claims (6)

An operation amount that is manipulated to change a control amount related to the air-fuel ratio of the internal combustion engine is set to a feedback correction value that is updated according to a difference between the actual value of the control amount and a target value, and the feedback correction value is set to “0”. In the air-fuel ratio control apparatus for an internal combustion engine that performs the feedback control of the air-fuel ratio by correcting with the feedback learning value updated to be close to
A feedback gain that variably sets the feedback gain in accordance with the degree of return so that the feedback gain of the feedback correction value decreases as the return from an overresponse of the feedback learned value proceeds when the feedback learned value is updated An air-fuel ratio control apparatus for an internal combustion engine, comprising: setting means. An operation amount that is manipulated to change a control amount related to the air-fuel ratio of the internal combustion engine is set to a feedback correction value that is updated according to a difference between the actual value of the control amount and a target value, and the feedback correction value is set to “0”. In the air-fuel ratio control apparatus for an internal combustion engine that performs the feedback control of the air-fuel ratio by correcting with the feedback learning value updated to be close to
When updating the feedback learning value, storage means for storing the value of the feedback learning value when the amount of deviation from the value at the start of the update is maximized;
A feedback gain that variably sets the feedback gain according to the deviation so that the feedback gain of the feedback correction value decreases as the deviation between the value stored by the storage means and the current value of the feedback learning value increases. Setting means;
An air-fuel ratio control apparatus for an internal combustion engine, comprising: The air-fuel ratio control device includes a main feedback control that feedback-adjusts the fuel injection amount to reduce the difference between the air-fuel ratio and the target air-fuel ratio, which are grasped from the output of the air-fuel ratio sensor that detects the exhaust oxygen concentration upstream of the catalyst. The feedback control of the air-fuel ratio is performed through sub-feedback control that feedback-adjusts the output of the air-fuel ratio sensor to reduce the deviation between the measured value of the exhaust gas concentration downstream of the catalyst and the target value thereof,
The air-fuel ratio control apparatus for an internal combustion engine according to claim 1 or 2, wherein the operation amount is an exhaust oxygen concentration downstream of the catalyst, and the operation amount is an output of the air-fuel ratio sensor. An operation amount that is manipulated to change a control amount related to the air-fuel ratio of the internal combustion engine is set to a feedback correction value that is updated according to a difference between the actual value of the control amount and a target value, and the feedback correction value is set to “0”. The air-fuel ratio feedback control is performed by correcting with the feedback learning value updated to be close to "
When updating the feedback learning value, the feedback gain is variably set in accordance with the degree of return so that the feedback gain of the feedback correction value decreases as the return from the excessive response of the feedback learning value proceeds. An air-fuel ratio control method for an internal combustion engine, characterized by comprising: An operation amount that is manipulated to change a control amount related to the air-fuel ratio of the internal combustion engine is set to a feedback correction value that is updated according to a difference between the actual value of the control amount and a target value, and the feedback correction value is set to “0”. The air-fuel ratio feedback control is performed by correcting with the feedback learning value updated to be close to "
When updating the feedback learning value, the value of the feedback learning value when the amount of deviation from the value at the start of the update becomes maximum is stored, and the stored value and the current value of the feedback learning value An air-fuel ratio control method for an internal combustion engine, wherein the feedback gain is variably set according to the deviation so that the feedback gain of the feedback correction value decreases as the deviation increases. The feedback control of the air-fuel ratio is a main feedback control that feedback-adjusts the fuel injection amount so as to reduce the difference between the air-fuel ratio and the target air-fuel ratio obtained from the output of the air-fuel ratio sensor that detects the exhaust gas oxygen concentration upstream of the catalyst. And sub-feedback control that feedback-adjusts the output of the air-fuel ratio sensor to reduce the deviation between the measured value of the exhaust oxygen concentration downstream of the catalyst and the target value,
The air-fuel ratio control method for an internal combustion engine according to claim 4 or 5, wherein the manipulated variable is an exhaust oxygen concentration downstream of the catalyst, and the manipulated variable is an output of the air-fuel ratio sensor.

Images