JP2002327634A - Control device of internal combustion engine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize EGR control and injection amount control excellent in responsiveness and precision. SOLUTION: An exhaust O2 concentration for every combustion in a cylinder is estimated by using an intake air amount signal, an intake air pressure signal, and instruction injection information. Specifically, by using an O2 amount in fresh air, and an O2 amount of EGR gas flowing into a manifold, an O2 amount in a gas flowing into the cylinder is obtained, and an O2 amount consumed by the instruction injection amount Qr is obtained. Then an O2 amount consumed by the instruction injection amount is subtracted from the O2 amount in the gas flowing into the cylinder to obtain the exhaust O2 amount in the cylinder. By using the exhaust O2 amount and the gas amount flowing into the cylinder, the exhaust O2 concentration is estimated. In this method, there is no generated time delay, nor chemical reaction delay of the O2 sensor itself until the exhaust gas arrives at an O2 sensor, compared with the case of detection of the exhaust O2 by the O2 sensor, whereby the exhaust O2 concentration can be estimated with excellent precision.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a control device for an internal combustion engine.

[0002]

2. Description of the Related Art In recent years, exhaust gas regulations for diesel engines have become increasingly strict, and the need for precise EGR control and injection quantity control has increased. For example, in EGR control, E
A method of measuring the amount of air sucked into a cylinder by an air flow meter provided in the intake system in order to increase the accuracy of the GR rate, and performing feedback control so as to match this to a target value (air flow meter F / B control) Has been mass-produced.

[0003]

However, in a transient state such as acceleration or deceleration, the amount of air actually entering the cylinder is different from the amount of air measured by an air flow meter. Various base controls have also been studied. This model-based control is based on an air flow signal measured by an air flow meter, and calculates a physical or transfer function of the air transmission delay based on the air flow signal, and estimates the air flow entering the cylinder in a transient state. .

[0004] However, the amount of EGR flowing back from the EGR path greatly changes due to the back pressure on the upstream side of the EGR (that is, the cylinder outlet side). In particular, in a turbo-equipped engine, since the back pressure changes greatly and transiently, the EGR amount changes, which results in an error in the EGR rate. In particular, in the engine with a variable turbo, the change in back pressure is large, and the conventional air flow meter F / B control and the model-based control based on it cannot control the EGR rate accurately, resulting in adverse effects such as generation of smoke. Could not be prevented.

[0005] In addition, regarding the injection amount control, in the case of a diesel engine, the injection amount differs from the command value due to a manufacturing tolerance of the fuel injection device, a change over time, or the like. Problems such as generation and insufficient torque occur. As an attempt to solve this, a method of installing an O ₂ sensor for detecting the exhaust O ₂ concentration in the exhaust system and performing feedback control of the fuel amount based on the exhaust O ₂ concentration has been studied. However, O ₂
Since the sensor is installed, a time delay occurs before the exhaust gas reaches the installation position. Also, due to the chemical reaction delay of the O ₂ sensor itself, the actual exhaust O ₂ concentration and O ₂
It is different from the exhaust O ₂ concentration detected by the _second sensor, control accuracy, particularly, control accuracy at the time of transition there is a problem that greatly reduced. The present invention has been made on the basis of the above circumstances, and its object is to provide a highly responsive and accurate EGR
Control and injection amount control.

[0006]

According to the present invention, there is provided an intake air amount signal output from an intake amount measuring means,
Using air pressure signal output from the intake pressure sensor, and the command injection quantity information calculated by the command injection amount calculating means predicts the exhaust O ₂ concentration of each combustion in the cylinder, the predicted exhaust O ₂ It is characterized in that at least one of the EGR valve and the fuel injection amount is controlled according to the concentration.

In this configuration, the total amount of gas flowing into the cylinder can be detected by using the intake pressure signal, and the new air amount (EGR gas) flowing into the intake passage is determined by the intake air amount signal and the intake pressure signal. Can be detected with high accuracy. Further, the EGR gas amount flowing into the intake passage can be obtained from the total gas amount and the fresh air amount flowing into the intake passage. Furthermore, by adding the command injection amount information, the exhaust O ₂ concentration after each injection can be accurately predicted. According to this method, the exhaust O ₂ concentration can be predicted before the O ₂ concentration of the exhaust gas is actually detected by the sensor. Therefore, when the exhaust O ₂ concentration is used for the injection amount control or the EGR control, the control can be performed with good response.

According to a second aspect of the present invention, in the control device for an internal combustion engine according to the first aspect, the exhaust O ₂ concentration predicting means includes:
Consumption O to calculate O ₂ amount consumed by command injection amount
_{It has two-} quantity calculation means. In this case, by calculating the O ₂ amount consumed by the command injection amount, the exhaust O ₂ concentration after the injection can be accurately calculated.

According to a third aspect of the present invention, in the control device for an internal combustion engine according to the first or second aspect, the exhaust O ₂ concentration predicting means calculates the O ₂ amount in the gas (including the EGR gas) flowing into the cylinder. , O ₂ in the fresh air sucked into the intake passage
It is calculated using the amount and the O ₂ amount in the EGR gas. As a result, the amount of O ₂ flowing into the cylinder can be calculated with high accuracy, so that the prediction accuracy of the exhaust O ₂ concentration is improved.

According to a fourth aspect of the present invention, in the control apparatus for an internal combustion engine according to the third aspect, the exhaust O ₂ concentration predicting means includes:
Using the previously calculated predicted value of the past exhaust O ₂ concentration, E
The amount of O ₂ in the GR gas is calculated. In this case, as compared with the case where the O ₂ concentration of the exhaust gas is actually detected by the sensor, there is no influence of the detection delay, so that the exhaust O ₂ concentration is accurately obtained.
_Two concentrations can be predicted.

According to a fifth aspect of the present invention, there is provided the control device for an internal combustion engine according to any one of the first to fourth aspects, further comprising an O ₂ sensor installed in the exhaust passage and detecting an actual exhaust O ₂ concentration. A learning function is provided so that the predicted value of the exhaust O ₂ concentration matches the output value of the O ₂ sensor in a steady state. Thereby, the error of the exhaust O ₂ concentration prediction means can be reduced, and the prediction accuracy of the exhaust O ₂ concentration is improved.

(6) A control device for an internal combustion engine according to any one of the above (1) to (4), wherein an O ₂ sensor installed in an exhaust passage for detecting an actual exhaust O ₂ concentration;
The exhaust O ₂ concentration of the exhaust gas discharged from the cylinder is O ₂
Filtering means for correcting the time delay until detection by the sensor, and learning correction so that the predicted value of the exhaust O ₂ concentration matches the output value of the O ₂ sensor after filtering is performed by the filtering means. And a learning function of calculating the amount and correcting the predicted value of the exhaust O ₂ concentration with the learning correction amount. According to this configuration, by performing the filtering, it is possible to accurately detect the difference between the predicted value of the exhausted O ₂ concentration and the measured value (the output value of the O ₂ sensor). As a result, accurate learning control can be realized even during transition.

According to a seventh aspect of the present invention, in the control device for an internal combustion engine according to the sixth aspect, the update of the learning correction amount is prohibited under the condition that the change rate of the predicted exhaust O ₂ concentration is a predetermined value or more. I do. Generally, learning is performed in a steady state. However, in a normal operation, an operating state in a steady state is limited, and particularly in a region such as a low speed and a high load, the steady state hardly occurs. Therefore, when learning up to the transition time, it is important to what extent the learning is permitted.

An error occurs in the filtering when the signal of the exhaust O ₂ concentration changes rapidly. Therefore,
If the rate of change of the predicted exhaust O ₂ concentration is higher than a predetermined value, the accuracy of filtering cannot be guaranteed. Therefore, by prohibiting the update of the learning correction amount, learning in a transient state within a predetermined range can be realized. It is. Incidentally, the predicted exhaust O ₂ concentration as referred to herein, may be an exhaust O ₂ concentration in the exhaust gas discharged from the cylinder, may be exhaust O ₂ concentration after filtering.

In the control apparatus for an internal combustion engine according to the present invention, the dead time and the time constant until the exhaust gas discharged from the cylinder reaches the O ₂ sensor are determined in a steady state. Is characterized by detecting from the output value of the O ₂ sensor when the injection amount is slightly changed. Dead time and the time constant is an element of the filtering is previously adapted for each operating condition of the internal combustion engine, because there is a variation in the internal combustion engine and the O ₂ sensor, it is desirable to be able to correct during operation. In contrast, according to the present invention,
By changing the injection amount minutely in the steady state, the exhaust O
₍₂₎ By minutely changing the concentration and detecting how late the O ₂ sensor responds, it is possible to improve the accuracy of the filtering. Thereby, accurate learning can be performed at the time of transition.

According to a ninth aspect of the present invention, in the control device for an internal combustion engine according to any one of the sixth to eighth aspects, a catalyst is provided in the exhaust passage, and an O ₂ sensor is provided downstream of the catalyst. It is characterized by the following. In recent years, the mounting of catalysts and aftertreatment systems such as NOx catalysts and diesel particulate filters has increased. The O ₂ sensor is susceptible to pressure, and when these post-processing systems are installed, the detection accuracy of the O ₂ sensor decreases because the catalyst or the like becomes a pressure loss body and the pressure near the sensor increases. There is a problem of doing. On the other hand, in the present invention, even if the O ₂ sensor is disposed downstream of the catalyst, the learning O ₂ concentration can be accurately corrected by using the filtering that accurately predicts the instantaneous exhaust O ₂ concentration every moment and corrects the delay. The output value of the O ₂ sensor can be used.

In the control device for an internal combustion engine according to any one of claims 6 to 9, when addition of fuel exhaust pipe for catalyst control or post injection is performed, learning correction is performed. It is characterized in that updating of the quantity is prohibited.
For the purpose of controlling the catalyst, techniques such as addition of fuel to the exhaust pipe for adding fuel to the exhaust pipe for a predetermined period and post-injection of injecting fuel into the cylinder after combustion has been developed have been developed.

However, these fuel components may lead to a decrease in the detection accuracy of the O ₂ sensor. In addition, when the catalyst temperature rises due to the activation of the chemical reaction in the catalyst due to the addition of fuel, and the soot in the catalyst is burned, the exhaust O ₂ concentration near the O ₂ sensor is discharged from the cylinder. From the exhaust O ₂ concentration. For this reason, it is difficult to ensure the accuracy of the learning control. Therefore, when these catalyst controls are performed, erroneous learning can be prevented by prohibiting updating of the learning correction amount.

[0019] In the control device of any one of the internal combustion engine according to claim 6-10 (unit of claim 11), when the absolute value of the learning correction amount is a predetermined value or more, it is determined that the O ₂ sensor abnormality It is characterized by the following. O ₂ sensors, in order to ensure the detection accuracy, it is necessary to keep the sensor element to a temperature as high as 600 to 800 ° C.. In addition, there is a possibility that a failure such as the sensor element being cracked due to being wetted. In these cases, there arises a problem of learning an incorrect value. Therefore, when the absolute value of the learning correction amount is equal to or larger than a predetermined value, it is determined that the O ₂ sensor is abnormal.

According to a twelfth aspect of the present invention, in the control device for an internal combustion engine according to the eleventh aspect, when it is determined that the O ₂ sensor is abnormal, the learning correction by the learning function is prohibited. Although the predicted exhaust O ₂ concentration includes a model error, a certain degree of accuracy is guaranteed. Therefore, even if the learning correction is prohibited due to the abnormality of the O ₂ sensor, the control can be realized with the model predicted value, so that it is possible to prevent the EGR control and the injection amount control using the exhaust O ₂ concentration from being greatly affected.

[0021] In the control device of any one of the internal combustion engine according to claim 1 to 12 (unit of claim 13) has a target value of the exhaust O ₂ concentration for each operating region of the engine, the exhaust O ₂
The EGR valve is feedback-controlled so that the predicted value of the concentration matches the target value. As a result, the exhaust O
_{(2) Since the} concentration can follow the target value, the emission is improved.

According to a fourteenth aspect of the present invention, in the control apparatus for an internal combustion engine according to any one of the first to twelfth aspects, the predicted value of the exhaust gas O ₂ concentration when calculated using the command injection amount is predetermined. The commanded injection amount is corrected by recalculating the injection amount so as to match the target value. This makes it possible to control the exhaust O ₂ concentration for each injection, thereby improving the controllability of the exhaust O ₂ concentration and improving the emission.

In the control device for an internal combustion engine according to any one of the first to twelfth aspects, the predicted value of the exhaust O ₂ concentration when calculated using the command injection amount is predetermined. The upper limit is set for the command injection amount so as not to exceed the limit value on the rich side. Thus, it is possible to prevent a large amount of smoke from occurring while minimizing the correction of the injection amount.

(Means of Claim 16) In the control device for an internal combustion engine according to claim 15, the limit value on the rich side is:
It is at least a function of the engine speed. Thereby, when the torque at the time of low rotation is required, the amount of correcting the injection amount can be reduced, so that the acceleration is improved.

[0025]

Next, an embodiment in which the present invention is applied to a diesel engine will be described with reference to the drawings. FIG.
1 is an overall configuration diagram illustrating a control system of a diesel engine 1. FIG. First, the overall configuration of the system will be described with reference to FIG. The present system is applied to a diesel engine (hereinafter abbreviated as engine 1) having an EGR function (described below) for returning a part of exhaust gas to intake air.

In the engine 1, high-pressure fuel stored in a common rail (not shown) is supplied from an injector 2 attached to a cylinder head of the engine 1 to a combustion chamber 1.
A common rail type injection system for injecting a. The EGR function includes an EGR passage 5 that connects the intake passage 3 and the exhaust passage 4, and an EGR valve 6 that adjusts an amount of exhaust gas (EGR amount) that recirculates through the EGR passage 5.

In the intake passage 3, an air flow meter 7 and a compressor 8A of a variable turbo 8 are provided upstream of a connection point with the EGR passage 5. A diesel throttle 9 (hereinafter referred to as a throttle 9) is provided downstream of the compressor 8A. I have it. Further, an intake pressure sensor 10 for detecting the intake pressure in the intake passage 3 and an intake air temperature sensor 11 for detecting the temperature of the air in the intake passage 3 are provided downstream of the throttle 9. In this embodiment, the upstream side of the throttle 9 in the intake passage 3 is called an intake pipe 3A,
The downstream side is referred to as a manifold 3B. In the exhaust passage 4, a variable turbo 8 is provided downstream of a connection point with the EGR passage 5.
Exhaust turbine 8B is provided, and the exhaust turbine 8B
Further downstream, an O ₂ sensor 12 for detecting the O ₂ concentration of the exhaust gas is attached.

The air flow meter 7 and the intake pressure sensor 1
Each information of the air system detected by the intake air temperature sensor 11 and the O ₂ sensor 12 is output to an electronic control unit (hereinafter referred to as an ECU 13) for controlling the operation of the present system. In addition to the above sensors, the present system includes a rotation angle sensor 1 that outputs a signal in synchronization with the rotation angle of the engine 1.
4. a water temperature sensor 15 for detecting a cooling water temperature of the engine 1;
An accelerator opening sensor 17 and the like for detecting the accelerator opening from the depression amount of the accelerator pedal 16 are provided, and various detected information is output to the ECU 13.

Next, a control method of the present system will be described. (First Embodiment) FIG. 2 is a block diagram showing the contents of control by the ECU 13, in which exhaust O ₂ concentration prediction calculation, target exhaust O ₂ concentration calculation, injection amount correction calculation, EGR operation amount calculation,
And five blocks for learning correction amount calculation. FIG.
3 is a drawing of an air system in which various symbols used for describing the control contents of the present system are shown. FIG. 4 (A),
3B is a flowchart illustrating a control procedure of the ECU 13.

The main routine shown in FIG. 4A is calculated for each injection cycle in synchronization with the injection (Ne synchronization). By performing the injection synchronization, the exhaust O ₂ concentration for each injection can be calculated, so that the accuracy is improved and the emission is improved. On the other hand, the main routine shown in FIG. 4B is calculated at time synchronization, for example, every 16 ms. This is because the responsiveness of the EGR valve 6 does not depend much on the engine speed Ne. However, there is no problem if the calculation (injection synchronous calculation) is performed after the routine shown in FIG. 4A (after Step 600).

First, the contents of the main routine shown in FIG. Step 100 (command injection amount calculation means): Calculates the command injection amount Qr. Although the flowchart showing the calculation procedure is omitted,
Generally, it is determined by the sum of a basic injection amount calculated from a map of the engine speed Ne and the accelerator opening and an injection amount for obtaining a driving force required for an air conditioner or the like. Step 200 (Exhaust O ₂ concentration prediction means): Predict the exhaust O ₂ concentration for each combustion in the cylinder.

Step 300... The predicted value of Step 200 will be described later.
The accuracy is improved by adding the learning correction amount calculated in Step 611. Step 400: Calculate the target exhaust O ₂ concentration. Step 500: The injection amount is corrected using the predicted exhaust O ₂ concentration after learning calculated in Step 300 and the target exhaust O ₂ concentration calculated in Step 400. Step 600: Error learning is performed on the predicted exhaust O ₂ concentration after the injection amount is corrected, based on the actually measured exhaust O ₂ concentration. On this occasion,
Performing the integral learning in the steady state improves the accuracy in the steady state.

Next, the contents of each step performed in the main routine of FIG. 4A will be described in detail. a) FIG. 5 is a subroutine showing the processing procedure of Step 200, and the detailed processing procedure for each step shown in FIG.
This is shown in each flowchart of FIG. Step 210: The amount of gas Mcld flowing into the cylinder is calculated. Specifically, as shown in FIG. 6, Steps 211 to 213... The intake pressure Pm detected by the intake pressure sensor 10, the intake temperature sensor 11
After sequentially reading the intake air temperature Tm and the rotational speed Ne detected in step 2, step 214: calculating the volumetric efficiency η as a function of Ne and Pm, and step 215: calculating the gas amount Mcld flowing into the cylinder from the gas state equation and η. calculate. Note that R used in the gas state equation is a gas constant and may be a constant value.

Step 220: Calculate the fresh air amount MDth flowing into the manifold 3B. Specifically, as shown in FIG. 7, Steps 221 to 223... Sequentially read the flow rate MAFM measured by the air flow meter 7, the intake pressure Pm, and the intake temperature Tm, and then calculate Step 224. Step 225: The fresh air amount MDth is calculated. Here, the pressure in the intake pipe 3A from the air flow meter 7 to the throttle 9 is substituted by Pm, the mass increase in the intake pipe 3A is calculated from the gas state equation, and the following equation is used. The fresh air amount MDth can be calculated from the law of conservation of mass. MAFM × (2 / cylinder number) −MDth = ΔP · VIN / (Tm · R)

Step 230: E flowing into the manifold 3B
The GR gas amount MEGR is calculated. Specifically, as shown in FIG. 8, Steps 231 to 235... Pm,
After sequentially reading Tm, ΔP, MDth, and Mcld,
Step 236: EGR gas amount MEGR is calculated. That is, the mass increase in the manifold 3B is calculated from the gas state equation,
The EGR gas amount MEGR is calculated from the law of conservation of mass in the manifold 3B expressed by the following equation. MDth + MEGR-Mcld = [Delta] P * Vm / (Tm * R) ...............

Step 240: O in the gas flowing into the cylinder
Calculating _two amount O _2-cld. Specifically, as shown in FIG.
, MEGR, and the corrected exhaust O ₂ concentration Cex-c (in ₁ )... (Corrected exhaust O ₂ concentration before n ₁ cycle) calculated in the past cycle are sequentially read, and Step 244.
-c (in ₁ ) is stored in the memory as the O ₂ concentration CEGR in the EGR gas.

Note that n ₁ takes into account the delay of gas inflow and may be a constant or a function of Ne. Briefly, C
ex-c may be annealed. In order to improve the accuracy, it is preferable that n ₁ is a function of the past EGR gas amount MEGR. Specifically, there is a method of multiplying the volume of the EGR passage 5 (EGR pipe) by the gas density and dividing the result by the EGR gas amount per cycle (past value).

Subsequently, Step 245: The O ₂ concentration Cm-IN at the place where the fresh air gas and the EGR gas are mixed is calculated, and
ep246 ... O ₂ concentration Cm-cld in the gas flowing into the cylinder
Is calculated. Here, n ₂ takes into account the delay of gas inflow and may be a constant or a function of Ne. Also, Cm-IN
May be processed. Then, Step 247... Cylinder inflow gas amount Mcld is read, and Step 248.
Calculate the cylinder inflow O ₂ amount O ₂ -cld by multiplying with ld.

Step 250: Model predicted exhaust O ₂ concentration Cex-m
Calculate dl. Specifically, as shown in FIG.
d, O _2- cld, and after that sequentially reads the command injection amount Qr, Step254 (consumed O ₂ amount calculating means) ... command injection amount Qr
Calculating the amount of O ₂ consumed by. K1 (constant)
Is the O ₂ mass consumed per unit fuel amount. However, the rate of incomplete combustion is large depending on the combustion mode (1% or more)
If so, it is better to fix it.

Further, Step 255: The model predicted exhaust O ₂ concentration Cex-mdl is calculated. Specifically, the amount of oxygen O _2-cld flowing into the cylinder by subtracting the consumed oxygen quantity O _2-qr calculates the residual oxygen quantity, which the gas quantity Mcld and fuel quantity K2 × Qr flowing into the cylinder Cex-mdl can be calculated by dividing by the sum. K2 is a constant of the fuel density.

B) FIG. 11 is a subroutine showing the processing procedure of Step 300. Here, the model predicted exhaust O ₂ concentration Cex-mdl calculated in Step 301... Step 255 is read, and further Step 302.
After reading the learning correction amount CLEARN calculated by the later-described Step611, calculates the predicted exhaust O ₂ concentration Cex-s after learning by adding CLEARN in Step303 ... Cex-mdl.

C) FIG. 12A is a subroutine showing the processing procedure of Step 400. Here, after reading Steps 401 to 402... Ne and QR, Step 403... The target exhaust gas O from the map shown in FIG.
_2. Calculate the concentration Cex-trg. For example, in the figure, Ne = N1
, Qr = Q1, .beta. Is calculated by a map search. Here, the target exhaust O ₂ concentration Cex-trg embedded in the map is set so that the O ₂ concentration discharged from the engine 1 is always good for each driving range so that good emission, fuel efficiency, drivability, etc. can be obtained. This is a value obtained in advance by experiments. Subsequently, β obtained in Step 404... Step 403 is stored in the memory as the target exhaust O ₂ concentration Cex-trg.

D) FIG. 13 is a subroutine showing the processing procedure of Step 500. Here, after reading the predicted exhaust O ₂ concentration Cex-s after learning and the target exhaust O ₂ concentration Cex-trg, Step 501 to 502.
tep503: Calculate the difference ΔCex between the two. Then, Step5
04… After inputting the command injection amount Qr, Step 505… ΔCex
Is greater than or equal to 0. If ΔCex> 0,
That is, if Cex-s is smaller than Cex-trg, Step 50
6 to 508: Correct the injection amount.

Specifically, Step 506: After inputting the amount of gas Mcld flowing into the cylinder, Step 507: Target exhaust O ₂
The injection correction amount ΔQ is calculated so as to match the density Cex-trg. The equation of Step 507 can be derived by eliminating the consumed oxygen amount O ₂ -qr from the equations of Step 254 and Step 255 and differentiating both sides by Qr under the assumption that MclddK2 × Qr. Next, Step 508: The corrected injection amount Qc is calculated from the command injection amount Qr and the injection correction amount ΔQ. Furthermore, the exhaust O ₂ concentration, the target exhaust O ₂ concentration since changing the injection amount to match the Cex-trg, Step509 ... target exhaust O ₂ concentration Cex-trg evacuated after injection quantity corrected O ₂ concentration C
Store in memory as ex-c.

On the other hand, if the above Step 505... ΔCex ≦ 0, Steps 510 to 512. here,
This is the same as Steps 506 to 508, but the injection amount is corrected to the increase side, so the correction amount is reduced by the coefficient α (= 0 to 1). This is to prevent the torque from being higher than required by the driver. The coefficient α is
Adapt beforehand from the relationship between emissions and drivability. If drivability is prioritized, α is small, and if emission is prioritized, α is large.

The processing of Steps 510 to 512 will be specifically described. Step 510: After inputting the gas amount Mcld flowing into the cylinder, Step 511: calculate the injection correction amount ΔQ, and Step 512: correct the injection amount Qc from the command injection amount Qr and the injection correction amount ΔQ.
Is calculated. Further, Step 513: Exhaust O ₂ after correcting the injection amount
The density Cex-c is stored in the memory. In this Step 513, the predicted exhaust O ₂ concentration Cex-s after the learning and the target exhaust O ₂ concentration Cex-t
It is a value that internally divides between rg with α.

E) FIG. 14 is a subroutine showing the processing procedure of Step 600. Here, Steps 601 to 604... After sequentially reading the exhaust gas O ₂ concentration Cex-c and the output values Cex-sensor Qc and Ne of the O ₂ sensor 12 after the correction of the injection amount calculated in Step 500,
Step 605: A learning area is determined. FIG. 15 shows an example of this learning area. Subsequently, a learning value (learning correction amount) is read from the learning area in Step 606, and Steps 607 to 609 are determined whether or not the apparatus is in a steady state. However, this is an example of determining a steady state.

FIG. 16 shows the processing of Steps 607 to 609.
This will be described with reference to FIG. Note that Qo, Qc, t shown in FIG.
o is a conforming value, for example, Qo = 3 mm ³ / st, Qc = 3 mm
³ / st, to = 3 [s], etc. Also, (60 / Ne) × (2 / cylinder number) in Step 608 is a time interval for each injection cycle, and KI in Step 610 is an integration constant (for example, 0.
05 etc.).

Step 607: Injection amount Qc-start at the start of determination
Therefore, the injection amount Qc is within a predetermined range (Qc-start-Qo to Qc-
start + Qo). Step 608: The counter tc is increased by a time interval for each injection cycle. Step 609: When tc is equal to or longer than the predetermined time to, it is determined to be in a steady state. If the determination result in Step 607 is NO, Step 612... Injection amount Qc is replaced with the injection amount Qc-start at the start of the determination, and then Step 613... Counter tc is reset (returned to 0) and the routine returns to the start.

Next, the contents of the main routine shown in FIG. 4B will be described. Step 1100: A basic EGRV operation amount (basic control current of the EGR valve) corresponding to the current operation state of the engine 1 is calculated. Step 1200: Calculate the exhaust O ₂ F / B correction amount (hereinafter referred to as the F / B correction amount) for the basic EGRV operation amount. Step 1300: A final EGRV operation amount is calculated by adding the F / B correction amount to the basic EGRV operation amount.

Next, the contents of each step performed in the main routine of FIG. 4B will be described in detail. a) FIG. 17 is a subroutine showing a processing procedure of Step 1100. Here, Steps 1101 to 1102... Ne and Qr are read, and then Step 1103... From the map shown in FIG.
The GRV operation amount is calculated. For example, in the figure, Ne = N2,
When Qr = Q2, .alpha. Is calculated by map search. Here, the basic EGRV operation amount embedded in the map is set in an experiment in advance so that emission, fuel efficiency, drivability, etc., are ideal values for each operating range of the engine 1 in the initial central product of the control system. This value is obtained by setting the EGR rate. Thereafter, α obtained in Step 1104... Step 1103 is stored in the memory as the basic EGRV operation amount IEBSE.

B) FIG. 18 is a flowchart showing the processing procedure of Step 1200.
Is a routine. Here, Steps 1201 to 1202._TwoConcentration Cex-trg
And predicted exhaust O after learning _TwoAfter reading the density Cex-s,
Step1203: Calculate the F / B amount IEO2FB. This calculation includes one
General proportional-integral control (PI control) and state quantity F / B control
Conceivable. c) FIG. 19 is a subroutine showing a processing procedure of Step 1300.
is there. Here, Step1301 ~ 1302 ... F / B amount IEO2FB and basic
After reading the EGRV operation amount IEBSE, Step1303 ...
To obtain the final EGRV operation amount IEFIN.

(Effect of the First Embodiment) In the above-described system, the exhaust O ₂ concentration for each combustion in the cylinder is calculated based on the intake air amount MAFM, the intake pressure Pm, and the command injection amount Qr. . According to this method, as compared with the case where the exhaust O ₂ concentration is detected by the O ₂ sensor 12 attached to the exhaust passage 4, a time delay that occurs until the exhaust gas reaches the O ₂ sensor 12, _Since there is no chemical reaction delay of the ₂ sensor 12 itself, the exhaust O ₂ concentration can be accurately predicted. Therefore, by performing EGR control or injection amount control in accordance with the predicted exhaust O ₂ concentration, the O ₂ sensor 1
As compared with the case where the injection amount control or the EGR control is performed based on the exhaust O ₂ concentration detected in step ₂ , the responsiveness is excellent, and the control accuracy, especially the control accuracy in a transient state, is greatly improved.

(Second Embodiment) FIG. 20 is a block diagram showing the contents of control by the ECU 13. In comparison with the first embodiment (FIG. 2), the target exhaust O _The difference is that the limiting exhaust O ₂ concentration is used instead of the ₂ concentration. This limit exhaust O ₂ concentration is a concentration that indicates the limit value because the smoke increases when the exhaust O ₂ concentration decreases during acceleration or the like. If the injection amount is controlled so as not to be on the rich side (ie, on the side where the exhaust O ₂ concentration is lower) than this limit exhaust O ₂ concentration, smoke can be reduced.

FIGS. 21A and 21B are flow charts showing the control procedure of the ECU 13, which is the first embodiment (FIG. 4).
Step 400A and Step 500A are different from those of FIG. Step
As shown in FIG. 22, 400A... The limit exhaust O ₂ concentration is obtained from a map using the rotation speed Ne and the allowable smoke as parameters. That is, since smoke is a function of the number of revolutions,
Assuming that the limit exhaust O ₂ concentration is a function of the number of revolutions, the accuracy is improved, and the injection amount can be injected up to the limit of the desired smoke concentration limit. FIG. 23 is an example of a map. This map is different depending on the engine shape, but does not change so much, so that it is easy to adapt.

Step 500A: This processing procedure (subroutine)
Is shown in the flowchart of FIG. Here, Steps 502A, 503A, and 509A are different points from the first embodiment (FIG. 13). That is, Step 502A... The limit exhaust O ₂ concentration Cex-guard obtained in Step 400A is read, and Step 503A... The limit exhaust O ₂ concentration Cex-guard and the predicted exhaust O ₂ concentration C after learning.
The deviation ΔCex from ex-s is calculated. Thereafter, if ΔCex is greater than 0, Step 509A: Limit exhaust O ₂ concentration Cex-gua
rd is stored in the memory as the exhaust O ₂ concentration Cex-c after the injection amount is corrected. If ΔCex ≦ 0, α (coefficient) = 0
And Thus, the injection amount is corrected only when the exhaust gas is richer than the limit exhaust O ₂ concentration.

FIG. 25 shows an operation example of the second embodiment.
Accelerator opening (1), command injection amount (2)
, Air flow meter flow rate (3), intake pressure (4), predicted exhaust O ₂ concentration after learning (5), corrected injection amount (6), and transient smoke (7). In this operation example, when the accelerator opening increases during acceleration or the like, (5) if the predicted exhaust O ₂ concentration after learning is on the rich side (lower exhaust O ₂ concentration) than the limit exhaust O ₂ concentration, (6) The injection amount is corrected and reduced from the injection amount before correction. As a result, (7) smoke during transition can be reduced as compared with the case where the injection amount is not corrected.

[0058] (Third Embodiment) This embodiment is an example having a filtering means for exhaust O ₂ concentration in the exhaust gas discharged from the cylinder to correct the time delay before it is detected by the O ₂ sensor 12 is there. FIG. 26 is a block diagram showing the control contents of the ECU 13. Control blocks 2000 and 2100 are added to the first embodiment, and
The content of the learning amount calculation shown in the control block 600 of the embodiment is changed and replaced with a control block 2200. Note that the control blocks 200 to 500 and 1100 to 1300 are the same as those in the first embodiment, and a description thereof will be omitted.

First, the processing contents of the control block 2000 (filtering) will be described with reference to the flowchart shown in FIG. Step 2001: The exhaust O ₂ concentration Cex-c after the injection amount correction calculated in the above-described Step 500 is read. Step2002: Read Qc and Ne. Step2003: A time constant TA and a dead time TB are obtained from a map of Qc and Ne. It should be noted that the time constant TA and the dead time TB, which are elements of the filtering, are adapted in advance for each operating condition of the engine 1. Here, since the adaptation is performed in a limited range from the cylinder to the O ₂ sensor 12, it is possible to adapt with high accuracy only by filtering using the time constant TA and the dead time TB.

Step 2004: Using the time constant TA and the dead time TB obtained in Step 2003, a time delay until the exhaust O ₂ concentration of the exhaust gas discharged from the cylinder is detected by the O ₂ sensor 12 is filtered. . Here, an example is shown in which filtering is performed based on the dead time and the first-order lag.
Note that fA in Step 2004 is a time constant TA and a dead time TB
Is a function for filtering by, for example, FIG.
It has the function shown in (b).

Next, a control block 2100 (learning inhibition determination)
Will be described based on the flowchart shown in FIG. Step 2101: Read the exhaust O ₂ concentration Cex-c after the injection amount is corrected. Step 2102: Differentiate the read exhaust O ₂ concentration Cex-c.
Differentiation can be performed by, for example, a difference such as dividing the amount of change within a predetermined time by the time. However, since the difference easily causes noise, the difference value may be smoothed and used.

Step 2103: It is determined whether or not the absolute value | ΔCex-c | of the differential value obtained in Step 2102 is larger than a predetermined value ΔCex-max. The predetermined value depends on the required filtering accuracy. For example, a value such as 5% / s. Step2104... | Cex-c | is larger than a predetermined value (the determination result is YES), so that the accuracy of the filtering cannot be guaranteed, and the learning inhibition flag is turned ON. Step2105 ... | ΔCex-c | is smaller than the predetermined value (the determination result is NO), the learning prohibition flag is turned off.

Next, the processing contents of the control block 2200 (learning amount calculation) will be described based on the flowchart shown in FIG. Steps 2201 to 2204: Cex-cc filtered in the above Step 2004, the output value Cex-sensor of the O ₂ sensor 12,
Qc and Ne are sequentially read. Step 2205: A learning area is determined from the map (see FIG. 15). Step 2206: A learning value (learning correction amount) CLEARN is read from the learning area.

Step 2207: A learning prohibition flag is determined based on the processing results of Steps 2104 and 2105. Step2208: When the learning prohibition flag is OFF (judgment result YES)
Is Cex-sensor, which is the measured value of exhaust O ₂ concentration, and Cex-cc after filtering (predicted exhaust O ₂
Density) to perform integral learning. Step2209… The value obtained by performing integral learning in Step2208 is the learned value CLEARN
Update as (X).

As described above, in this embodiment, since the learning value is calculated from the comparison between the predicted exhaust O ₂ concentration after filtering and the output value of the O ₂ sensor, the learning control can be performed with high accuracy even in the transient state. Can be realized. However, an error occurs in the filtering when the signal of the exhaust O ₂ concentration changes rapidly. Therefore, the predicted exhaust O ₂ concentration change rate is equal to or higher than a predetermined condition (in the judgment of Step2103 | ΔCex-c | is larger than a predetermined value), the can not guarantee the filtering accuracy, prohibits updating of the learning correction amount Thus, learning in a transient state can be realized within a predetermined range.

(Fourth Embodiment) In this embodiment, the cylinder
Exhausted exhaust gas is O_TwoUntil it reaches the sensor 12
Dead time and time constant, small injection amount in steady state
O when changed _TwoDetect from output value of sensor 12
This is an example showing the method, and the processing procedure is shown in FIG.
Shown in the chart. Step 3000: Perform steady-state determination. This is, for example, | ΔC
ex-c | is less than or equal to a predetermined value A for a predetermined time t0
(See FIG. 30 (b)). In addition, within a predetermined time
The change in the injection amount is within the specified value and the change in the
A method such as within a fixed value may be used.

Step 3001: The injection amount is increased by a very small amount (see FIG. 30 (c)). Since increasing often the effect on torque exits, for example when a 1 mm ³ / st. Step 3002: An elapsed time t1 from when the injection amount is increased to when | ΔCex-c | reaches the predetermined value B is detected (FIG. 30 (d)).
reference). The time t1 is a dead time until a change in the exhaust O ₂ concentration in the cylinder due to a change in the injection amount is detected by the O ₂ sensor 12.

Step 3003: After increasing the injection amount, | ΔCex
The elapsed time t2 until -c | reaches the predetermined value C is detected (see FIG. 30 (e)). Step 3004: Correct the values of the time constant map and the dead time map which have been adapted in advance from t2 and (t2-t1). Although the correction method is omitted, for example, a method in which the ratio between the value of the dead time map and t1 is used as the correction coefficient is considered. Note that (t2-t1)
Is for detecting the dead time in the O ₂ sensor 12.

The dead time and the time constant, which are elements of the filtering, are adapted in advance for each operating condition of the engine 1. However, since there are variations in the engine 1 and the O ₂ sensor 12, it is desirable that they can be corrected during the operation. In contrast,
According to the method of the present embodiment, the exhaust O ₂ concentration is minutely changed by minutely changing the injection amount in the steady state,
_Since it is possible to detect how late the _two sensors 12 show a response, the accuracy of the filtering shown in the third embodiment can be improved, and accurate learning can be performed during a transition.

(Fifth Embodiment) This embodiment is shown in FIG.
In the exhaust passage 4, the catalyst 18 or the diesel
Is provided downstream of the catalyst 18.
To O_TwoThis is an example in which the sensor 12 is installed. Under catalyst 18
O on the downstream side_TwoWhen the sensor 12 is installed, the catalyst 18
And the pressure near the sensor rises._TwoC
There is a problem that the detection accuracy of the sensor 12 is reduced. this
On the other hand, in this embodiment, the exhaust O_TwoDark
Predict the degree, O from the cylinder_TwoDelay to sensor 12
By performing the filtering to correct, the catalyst 1
8 downstream of O_TwoEven if the sensor 12 is arranged,
O for learning correction _TwoThe output value of the sensor 12 can be used.
You.

(Sixth Embodiment) This embodiment is an example of the case where post-injection is performed for catalyst control, and the processing procedure is shown in the flowchart of FIG. The processing of this embodiment is performed before Step 2101 shown in FIG. 28 (between the start and Step 2101). Step 3101: It is determined whether a predetermined time has elapsed after the post injection is performed. Step3102: When a predetermined time has elapsed (judgment result YES)
Turns off the learning prohibition flag. This is because the unburned fuel remaining in the catalyst may continue to oxidize even after the post-injection ends, so the learning inhibition flag is cleared (OFF).
Is performed after a predetermined time or more after the end of post-injection, the learning accuracy is improved.

Step 3103: If the predetermined time has not elapsed (determination result NO), the learning inhibition flag is turned ON. The present embodiment can be applied to a case where fuel exhaust pipe addition is performed instead of post injection. In the fuel exhaust pipe addition, as shown in FIG. 31, fuel is added into the exhaust passage 4 from a fuel addition valve 19 installed in the exhaust passage 4.

As described above, when the post injection or the addition of the fuel exhaust pipe is performed, these fuel components are changed to the O ₂ sensor 12.
May lead to a decrease in detection accuracy of Further, the fuel reaction activates the chemical reaction in the catalyst 18 to raise the catalyst temperature, and when the soot in the catalyst 18 is burned, the exhausted O ₂ concentration near the O ₂ sensor 12 is discharged from the cylinder. It differs from the exhaust O ₂ concentration at the time. For this reason, it is difficult to ensure the accuracy of the learning control. Therefore, when these catalyst controls are performed, erroneous learning can be prevented by prohibiting updating of the learning value (learning correction amount) until a predetermined time has elapsed.

[0074] (Seventh Embodiment) This embodiment is based on the absolute value of the learning value to determine the abnormality of the O ₂ sensor 12, when it is determined that the abnormality is an example of a case of prohibiting the learning correction ,
The processing procedure is shown in the flowchart of FIG. In addition,
The processing of this embodiment is performed after Step 611 shown in FIG.

Step 3201... It is determined whether or not the absolute value of the learning value updated in Step 611 is larger than a predetermined value CERROR-MAX. Step3202 ... If the absolute value of the learning value is greater than a predetermined value (determination result YES), the ON the O ₂ sensor abnormality flag. Step3203: The learning value CLEARN (X) is set to zero so that the learning amount is not reflected. This may be such that the learning amount is not reflected in the control block 300 shown in FIG.

The O ₂ sensor 12 needs to maintain the sensor element at a high temperature of 600 to 800 ° C. in order to secure detection accuracy. Further, there is a possibility that a failure such as the sensor element being cracked due to being wetted. In these cases, there is a problem that an incorrect value is learned. Therefore, when the absolute value of the learning value is greater than a predetermined value, prohibits the learning correction is determined that abnormality in the O ₂ sensor 12. However, although the predicted exhaust O ₂ concentration includes a model error, a certain degree of accuracy is guaranteed. Therefore, even if the learning correction is prohibited due to the abnormality of the O ₂ sensor 12, the control can be realized with the model predicted value, so that it is possible to prevent the EGR control and the injection amount control using the exhaust O ₂ concentration from being significantly affected. .

In the above embodiment, the present invention is applied to the diesel engine 1, but can also be applied to a gasoline engine having an EGR function.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram showing a control system for a diesel engine.

FIG. 2 is a block diagram showing control contents by an ECU (first embodiment).

FIG. 3 is a drawing of an air system in which various symbols used to explain the control contents of the present system are shown.

FIG. 4 is a flowchart showing a control procedure of an ECU (first embodiment).

FIG. 5 is a flowchart showing a procedure for obtaining an exhaust O ₂ concentration.

FIG. 6 is a flowchart illustrating a calculation procedure of a cylinder inflow gas amount.

FIG. 7 is a flowchart showing a procedure for calculating a manifold inflow fresh air amount.

FIG. 8 is a flowchart showing a procedure for calculating the amount of EGR gas flowing into the manifold.

FIG. 9 is a flowchart showing a procedure for calculating a cylinder inflow O ₂ amount.

FIG. 10 is a flowchart showing a calculation procedure of exhaust O ₂ concentration.

FIG. 11 is a flowchart showing a procedure for obtaining a predicted exhaust O ₂ concentration after learning.

FIG. 12 is a flowchart (a) and a map (b) showing a procedure for obtaining a target exhaust O ₂ concentration.

FIG. 13 is a flowchart showing a procedure for correcting an injection amount.

FIG. 14 is a flowchart showing a procedure for learning and calculating a model error.

FIG. 15 is a map showing a learning area.

FIG. 16 is an explanatory diagram of a steady state.

FIG. 17 is a flowchart (a) and a map (b) showing a procedure for calculating a basic EGRV operation amount.

FIG. 18 is a flowchart showing a procedure for calculating an exhaust O ₂ F / B correction amount.

FIG. 19 is a flowchart showing a procedure for calculating a final EGRV operation amount.

FIG. 20 is a block diagram illustrating control by an ECU (second embodiment).

FIG. 21 is a flowchart illustrating a control procedure of an ECU (second embodiment).

FIG. 22 is a diagram schematically illustrating a method for obtaining a limit exhaust O ₂ concentration.

FIG. 23 is a map showing a correlation between a limit exhaust O ₂ concentration and a rotation speed.

FIG. 24 is a flowchart showing a procedure for correcting an injection amount.

FIG. 25 is a time chart showing an operation example of the second embodiment.

FIG. 26 is a block diagram illustrating control by an ECU (third embodiment).

FIG. 27 is a flowchart illustrating a filtering procedure (third embodiment);

FIG. 28 is a flowchart illustrating a processing procedure of learning prohibition determination (third embodiment).

FIG. 29 is a flowchart illustrating a processing procedure of learning amount calculation (third embodiment);

FIG. 30 is a flowchart (a) showing a processing procedure for obtaining a dead time and a time constant, and explanatory diagrams (b) to (e) (fourth embodiment).

FIG. 31 is an overall configuration diagram showing a control system for a diesel engine (fifth and sixth embodiments).

FIG. 32 is a flowchart illustrating a processing procedure when post-injection is performed (sixth embodiment).

FIG. 33 is a flowchart showing a processing procedure when learning correction is prohibited when it is determined that the O ₂ sensor is abnormal (seventh embodiment).

[Description of Signs] 1 Diesel engine (internal combustion engine) 3 Intake passage 3A Intake pipe (intake passage) 3B Manifold (intake passage) 4 Exhaust passage 5 EGR passage (EGR function) 6 EGR valve (EGR function) 7 Air flow meter (intake) The amount measuring means) 10 intake pressure sensor 12 O ₂ sensor 13 ECU (controller) 18 catalyst

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) F02D 41/22 380 F02D 41/22 380K 41/38 41/38 B 41/40 41/40 H 43/00 301 43/00 301G 301N 301W 45/00 320 45/00 320A 368 368G 368H F02M 25/07 550 F02M 25/07 550F 550R (72) Inventor Toshio Kondo 1-chome, Showa-cho, Kariya, Aichi Prefecture DENSO Corporation (72) Inventor Sumiko Kodaira 1-1-1, Showa-cho, Kariya-shi, Aichi F-term in DENSO Corporation (reference) 3G062 AA01 AA03 AA05 BA04 CA06 DA01 DA02 EA10 ED01 ED04 ED10 FA02 FA04 FA05 FA12 FA15 GA01 GA02 GA04 GA06 GA08 GA12 GA17 3G084 AA00 AA01 BA13 BA20 DA05 EA07 EA11 EB08 EB17 FA02 FA07 FA10 FA11 FA29 3G091 AA02 AA10 AA11 AA17 AA18 AA28 AB01 AB13 BA27 BA31 CA18 CB02 CB03 CB07 CB08 DA01 DA02 DB07 DB08 DB09 DB10 DB13 DB16 DC02 EA01 EA05 EA06 EA07 EA15 EA16 EA31 EA34 FB10 FC02 FC04 HA37 HB05 HB06 ECB13 EC07 ECA HA01Z HA04Z HA05Z HD05X HD05Z HE01Z HF08Z 3G301 HA02 HA13 JA12 JB01 JB09 LA00 MA12 MA23 MA26 NA08 NB07 NC02 ND01 ND21 ND45 NE17 NE23 PA01Z PA07Z PA10Z PD04A PD04B PD04Z PE01Z PF03Z

Claims (16)

[Claims] 1. A control device for an internal combustion engine having an EGR function for recirculating a part of exhaust gas to intake air, comprising: an intake air amount measuring means for measuring an amount of intake air taken into an intake passage; An intake pressure sensor for detecting an intake pressure of the internal combustion engine; a command injection amount calculation means for calculating a command injection amount based on at least an operation state of the internal combustion engine; an intake air amount signal output from the intake amount measurement means; intake pressure signal outputted from the pressure sensor, and using a command injection amount information calculated by the command injection amount calculating means, the exhaust O ₂ concentration estimating means for estimating the exhaust O ₂ concentration of each combustion in the cylinder An internal combustion engine, wherein at least one of an EGR valve and a fuel injection amount provided in the EGR function is controlled in accordance with the predicted exhaust O ₂ concentration predicted by the exhaust O ₂ concentration predicting means. Engine control device. 2. The control device for an internal combustion engine according to claim 1, wherein said exhaust O ₂ concentration predicting means has a consumed O ₂ amount calculating means for calculating an O ₂ amount consumed by said command injection amount. A control device for an internal combustion engine, comprising: 3. The control device for an internal combustion engine according to claim 1, wherein said exhaust O ₂ concentration predicting means calculates an O ₂ amount in gas (including EGR gas) flowing into said cylinder by said intake air. A control device for an internal combustion engine, wherein the control is performed using an O ₂ amount in fresh air and an O ₂ amount in EGR gas sucked into a passage. 4. A control apparatus for an internal combustion engine according to claim 3, wherein the exhaust O ₂ concentration prediction means, O ₂ in the EGR gas using the predicted value of the previous past the exhaust O ₂ concentration calculated A control device for an internal combustion engine, which calculates an amount. 5. A control device for an internal combustion engine according to claim 1, wherein said control device is provided in an exhaust passage and detects an actual exhaust O ₂ concentration.
Has _two sensors, the predicted value of the exhaust O ₂ concentration, so as to match the output value of the O ₂ sensor at the steady state, the control apparatus for an internal combustion engine, characterized in that gave a learning function. 6. The control device for an internal combustion engine according to claim 1, wherein said control device is provided in an exhaust passage and detects an actual exhaust O ₂ concentration.
₂ sensors, filtering means for correcting a time delay until exhaust O ₂ concentration of exhaust gas discharged from the cylinder is detected by the O ₂ sensor, and after performing filtering by the filtering means, A learning function of calculating a learning correction amount so that the predicted value of the exhaust O ₂ concentration matches the output value of the O ₂ sensor, and correcting the predicted value of the exhaust O ₂ concentration with the learning correction amount; A control device for an internal combustion engine, comprising: 7. The control device for an internal combustion engine according to claim 6, wherein the change rate of the predicted exhaust O ₂ concentration is not less than a predetermined value.
A control device for an internal combustion engine, wherein updating of the learning correction amount is prohibited. 8. A control device for an internal combustion engine according to claim 6, wherein a dead time and a time constant until exhaust gas discharged from said cylinder reaches said O ₂ sensor are calculated in an injection amount in a steady state. A control device for an internal combustion engine, wherein the control is performed based on an output value of the O ₂ sensor when the value is slightly changed. 9. A control device for an internal combustion engine according to claim 6, wherein a catalyst is provided in said exhaust passage, and said O ₂ sensor is provided downstream of said catalyst. Control device for an internal combustion engine. 10. The control device for an internal combustion engine according to claim 6, wherein the addition of the fuel exhaust pipe or the post-injection for controlling the catalyst updates the learning correction amount. A control device for an internal combustion engine, which is prohibited. 11. The control device for an internal combustion engine according to claim 6, wherein when the absolute value of the learning correction amount is equal to or greater than a predetermined value, the O _2.
A control device for an internal combustion engine, which determines that a sensor is abnormal. 12. A control device for an internal combustion engine according to claim 11, wherein when it is determined that said O ₂ sensor is abnormal, learning correction by said learning function is prohibited. . 13. The control device for an internal combustion engine according to claim 1, wherein a target value of the exhaust O ₂ concentration is provided for each operating region of the internal combustion engine, and the predicted value of the exhaust O ₂ concentration is A control device for an internal combustion engine, wherein feedback control of the EGR valve is performed so as to coincide with the target value. 14. The control device for an internal combustion engine according to claim 1, wherein the predicted value of the exhaust gas O ₂ concentration when calculated using the command injection amount is a predetermined target value. A control device for an internal combustion engine, wherein the command injection amount is corrected by recalculating the injection amount so as to coincide with the following. 15. The control device for an internal combustion engine according to claim 1, wherein the predicted value of the exhaust gas O ₂ concentration when calculated using the command injection amount is a predetermined rich side. A control device for an internal combustion engine, wherein an upper limit value is set for the command injection amount so as not to exceed a limit value of the following. 16. The control device for an internal combustion engine according to claim 15, wherein said rich limit value is at least a function of a rotation speed of said internal combustion engine.