JP2007231840A - Control device for internal combustion engine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a control device for an internal combustion engine controlling an internal combustion engine by appropriately grasping response delay of an air flow meter. <P>SOLUTION: This control device is provided with the air flow meter 41 provided in an engine intake air passage, a throttle model M10 estimating throttle passing air flow rate, and an AFM model M50 calculating estimated output value of the air flow meter by using AFM mode calculation formula based on estimation value of throttle passing air flow rate calculated by the throttle model, and controls the internal combustion engine by using actual measurement value and predicted output value of the air flow meter. Moreover, the control device is provided with an error calculation means calculating error between predicted output value of the air flow meter and actual measurement value of the air flow meter, and a model calculation formula correction means correcting the AMF model calculation formula based on error calculated by the error calculation means when throttle passing air flow rate increases and error calculated by the error calculation means when throttle passing air flow rate decreases. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

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

  The intake system of the internal combustion engine is divided into elements such as a throttle valve, an intake pipe, an intake valve, an air flow meter, etc., modeled for each element, and expressed by a mathematical expression to obtain an in-cylinder charged air amount of the internal combustion engine by calculation, 2. Description of the Related Art A control device for an internal combustion engine that controls the internal combustion engine by using the obtained in-cylinder charged air amount is known.

  Many of control devices using such an air model utilize the output of an air flow meter. However, the output of the air flow meter is delayed with respect to the air flow rate (hereinafter referred to as “air flow passage air flow rate”) that passes through the intake pipe provided with the air flow meter (for example, a hot wire type air flow meter). Is due to the heat mass of the air flow meter itself). For this reason, although the airflow meter can detect the airflow passing air flow rate relatively accurately during the steady operation of the internal combustion engine, the airflow meter cannot accurately detect the airflow passing air flow rate during the transient operation of the internal combustion engine.

  As described above, since the air flow meter has a response delay with respect to the actual air flow passing air flow rate, the air model uses the output of the air flow meter as the actual air flow passing air flow rate during the transient operation of the internal combustion engine, and the cylinder filling air When the amount is calculated, the calculated in-cylinder charged air amount becomes a value significantly different from the actual in-cylinder charged air amount, and as a result, the operation of the internal combustion engine cannot be optimally controlled. .

  Therefore, in the apparatus described in Patent Document 1, a response delay compensation element that compensates the response delay of the output of the air flow meter by phase advance compensation is provided on the input side of the intake system model, and the output of this response delay compensation element is used as the intake system model. I am going to enter it. That is, a value obtained by compensating the response delay existing in the air flow meter from the output value of the air flow meter using a mathematical expression using a time constant is input to another intake system model. Thus, the problem of the response delay of the air flow meter during transient operation of the internal combustion engine is improved.

JP 2003-314347 A JP2002-130042 JP 2000-320391

  By the way, in order to appropriately compensate for the response delay of the air flow meter, it is necessary to set the time constant used in the mathematical formula used for compensation to an optimum value. Here, in general, in order to set the time constant to an optimum value, an input value (for example, an output value of an air flow meter) and an output value (for example, an actual air flow passing air flow rate) used for the above compensation are used. ) Is required.

  However, since it is impossible to determine the actual airflow passing air flow rate, the relationship between the output value of the airflow meter as described above and the actual airflow passing air flow rate cannot be determined. Cannot be set. Therefore, in the apparatus described in Patent Document 1, the engine speed is used instead of the airflow passing air flow rate, and therefore the time constant is set based on the engine speed and the output change amount of the airflow meter. However, since the engine speed does not accurately represent the airflow passing air flow rate, it is difficult to set the time constant appropriately even if the time constant is set based on the engine speed.

  Therefore, an object of the present invention is to provide a control device for an internal combustion engine that appropriately grasps the response delay of the air flow meter and controls the internal combustion engine in consideration of the grasped response delay of the air flow meter.

In order to solve the above problems, in the first invention, the air flow meter provided in the engine intake passage, the throttle passing air flow rate estimating means for estimating the throttle passing air flow rate, and the throttle passing air flow rate estimating means are estimated. An air flow meter model for calculating an output value that would be output by the air flow meter using an air flow meter model calculation formula based on the estimated value of the air flow rate through the throttle, and the actual measurement of the air flow meter. In the control device for an internal combustion engine that controls the internal combustion engine using the value and the predicted output value of the air flow meter calculated by the air flow meter model, the predicted output value of the air flow meter calculated by the air flow meter model and the above An error calculating means for calculating an error between the measured value of the air flow meter and The air flow meter model calculation based on the error calculated by the error calculation means when the throttle passage air flow rate is increasing and the error calculated by the error calculation means when the throttle passage air flow rate is decreasing. Model calculation formula correcting means for correcting the formula is further provided.
Between the expected output value of the air flow meter when the throttle air flow rate is increasing and the actual measurement value, and between the expected output value of the air flow meter when the throttle air flow rate is decreasing and the actual measurement value It can be said that the airflow model calculation formula represents the response delay in the actual airflow meter relatively appropriately. According to the first aspect, since the air flow meter model calculation formula is corrected based on these errors, the air flow meter model calculation formula can be appropriately corrected.
Note that the predicted output value of the air flow meter calculated by the air flow meter model includes a corrected output value in addition to the predicted output value in the embodiment of the present invention described below.

  In the second invention, in the first invention, the model calculation formula correcting means is equal when the error calculated by the error calculating means is increasing and when the throttle passage air flow rate is decreasing. Thus, the above air flow meter model calculation formula is corrected.

In order to solve the above-mentioned problem, in the third invention, the air flow meter provided in the engine intake passage, the throttle passage air flow estimation means for estimating the throttle passage air flow, and the throttle passage air flow estimation means are estimated. An air flow meter model for calculating an output value that would be output by the air flow meter using an air flow meter model calculation formula based on the estimated value of the air flow rate through the throttle, and the actual measurement of the air flow meter. In the control device for an internal combustion engine that controls the internal combustion engine using the value and the predicted output value of the air flow meter calculated by the air flow meter model, the predicted output value of the air flow meter calculated by the air flow meter model and the above An error calculating means for calculating an error between the measured value of the air flow meter and The air flow meter model calculation formula is corrected based on the error calculated by the error calculation means when the internal combustion engine is accelerating and the error calculated by the error calculation means when the internal combustion engine is decelerating. Model calculation formula correcting means is further provided.
There is an error between the expected output value of the air flow meter when the internal combustion engine is accelerating and the actual measured value, and an error between the expected output value of the air flow meter when the internal combustion engine is decelerating and the actual measured value. If they are equal, it can be said that the airflow model calculation formula relatively appropriately represents the response delay in the actual airflow meter. According to the third aspect, since the air flow meter model calculation formula is corrected based on these errors, the air flow meter model calculation formula can be appropriately corrected.

  According to a fourth aspect, in the third aspect, the model calculation formula correcting means is configured such that the error calculated by the error calculating means is equal when the internal combustion engine is accelerating and when the internal combustion engine is decelerating. Correct the above air flow meter model formula.

  In a fifth invention, in any one of the first to fourth inventions, a time constant relating to a response time of the air flow meter with respect to a change in an actual throttle passage air flow rate is used in the air flow meter model calculation formula, The model calculation formula correcting means corrects the time constant.

  According to the present invention, it is possible to appropriately correct the air flow meter model calculation formula representing the response delay of the air flow meter, so that the response delay of the air flow meter can be properly grasped.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. The engine body 1 schematically shown in FIG. 1 represents a direct injection spark ignition type internal combustion engine. However, the present invention may be applied to another spark ignition internal combustion engine such as a port injection type spark ignition internal combustion engine or a compression self-ignition internal combustion engine.

  As shown in FIG. 1, in the embodiment of the present invention, the engine body 1 includes a cylinder block 2, a piston 3 that reciprocates within the cylinder block 2, and a cylinder head 4 fixed on the cylinder block 2. . A combustion chamber 5 is formed between the piston 3 and the cylinder head 4. In the cylinder head 4, an intake valve 6, an intake port 7, an exhaust valve 8, and an exhaust port 9 are arranged for each cylinder. Further, as shown in FIG. 1, a spark plug 10 is disposed at the center of the inner wall surface of the cylinder head 4, and a fuel injection valve 11 is disposed around the inner wall surface of the cylinder head 4. Further, a cavity 12 extending from the lower side of the fuel injection valve 11 to the lower side of the spark plug 10 is formed on the top surface of the piston 3. Further, the cylinder head 4 is provided with an intake valve control device 13 capable of continuously changing the phase angle and the valve lift amount of the intake valve 6.

  The intake port 7 of each cylinder is connected to a surge tank 15 via an intake branch pipe 14, and the surge tank 15 is connected to an air cleaner 17 via an intake pipe 16. A throttle valve 19 driven by a step motor 18 is disposed in the intake pipe 16. On the other hand, the exhaust port 9 of each cylinder is connected to an exhaust pipe 20, and the exhaust pipe 20 is connected to a casing 22 containing an exhaust purification catalyst 21. In the following description, portions of the intake branch pipe 14, the surge tank 15, the intake pipe 16 and the like from the throttle valve 19 to the intake valve 6 are referred to as an intake pipe portion 23.

  The electronic control unit (ECU) 31 comprises a digital computer, and is connected to each other via a bidirectional bus 32, a RAM (random access memory) 33, a ROM (read only memory) 34, a CPU (microprocessor) 35, an input. A port 36 and an output port 37 are provided. The surge tank 13 is provided with an intake pipe pressure sensor 40 for detecting the pressure of air (intake gas) in the intake pipe portion 23, and the intake pipe pressure sensor 40 is proportional to the pressure in the intake pipe portion 23. The output voltage is generated, and this output voltage is input to the input port 36 via the corresponding AD converter 38.

  The intake pipe 16 upstream of the throttle valve 19 is provided with an air flow meter 41 for detecting the flow rate of intake air flowing through the intake pipe 16, and the configuration of the air flow meter 41 in this embodiment will be described later. To do. Further, an intake air temperature sensor 42 for detecting the intake air temperature and an atmospheric pressure sensor 43 for detecting the atmospheric pressure are provided in the vicinity of the air cleaner 17. The throttle valve 19 is provided with a throttle valve opening sensor 44 that detects the opening of the throttle valve 19, and the throttle valve opening sensor 44 generates an output signal corresponding to the throttle valve opening. These air flow meter 41, intake air temperature sensor 42, atmospheric pressure sensor 43 and throttle valve opening sensor 44 output signals corresponding to intake air flow rate (mass flow rate), intake air temperature (atmospheric temperature), atmospheric pressure and throttle valve opening, respectively. This output signal is input to the input port 36 via the corresponding AD converter 38.

  The accelerator pedal 45 is connected to a depression amount sensor 46 that generates an output voltage proportional to the depression amount of the accelerator pedal 45, and the output voltage of the depression amount sensor 46 is input to the input port 36 via the corresponding AD converter 38. Entered. The crank angle sensor 47 generates an output pulse every time the crankshaft rotates 15 degrees, for example, and this output pulse is input to the input port 36. The CPU 35 calculates the engine speed from the output pulse of the crank angle sensor 47.

  On the other hand, the output port 37 is connected to the spark plug 10, the fuel injection valve 11, the intake valve control device 13, and the step motor 18 via a corresponding drive circuit 39.

  FIG. 2 is a schematic perspective view of an air flow meter 41 used in the present embodiment, and FIG. 3 is an enlarged perspective view of a heat ray measuring unit of the air flow meter 41 shown in FIG. In the present embodiment, the air flow meter 41 is a hot-wire flow meter, and as shown in FIG. 2, the air flow meter 41 is bypassed by a bypass passage that bypasses a part of the air flowing in the intake pipe 16 and the bypass passage. It has a hot-wire measuring unit 41a that measures the mass flow rate of the intake air, and a signal processing unit 41b that outputs a voltage corresponding to the measured mass flow rate. As shown in FIG. 3, the hot wire measuring unit 41a includes an intake air temperature measurement resistor 41a1 made of platinum heat wire, a support unit 41a2 that holds the intake air temperature measurement resistor 41a1 connected to the signal processing unit 41b, and a heating resistor. (Bobbin portion) 41a3 and a support portion 41a4 for connecting and holding the heating resistor 41a3 to the signal processing portion 41b. The signal processing unit 41b has a bridge circuit composed of an intake air temperature measurement resistor 41a1 and a heating resistor 41a3, and a temperature difference between the intake air temperature measurement resistor 41a1 and the heating resistor 41a3 is always constant by this bridge circuit. The power supplied to the heating resistor 41a3 is adjusted so that the power is maintained, and the supplied power is converted into a voltage and output. The relationship between the output voltage Vg of the air flow meter 41 and the flow rate of air passing through the intake pipe 16 in which the air flow meter 41 is disposed (hereinafter referred to as “air flow passage air flow rate”) is as shown in FIG.

  By the way, in the control device for the internal combustion engine, the combustion chamber 5 is filled when the intake valve 6 is closed in order to set the air-fuel ratio of the air-fuel mixture burned in the combustion chamber 5 of the internal combustion engine to the target air-fuel ratio. The amount of air (hereinafter referred to as “cylinder charged air amount Mc”) is estimated, and the fuel injection valve 11 is set so that the air-fuel ratio of the air-fuel mixture becomes the target air-fuel ratio based on the estimated cylinder charged air amount Mc. Defines the amount of fuel injected into the combustion chamber 5 (or intake passage) of the internal combustion engine (hereinafter referred to as “fuel injection amount”). Therefore, in order to accurately set the air-fuel ratio of the air-fuel mixture combusted in the combustion chamber 5 of the internal combustion engine to the target air-fuel ratio, it is necessary to accurately estimate the cylinder charge air amount Mc.

  Usually, the in-cylinder charged air amount Mc is estimated from a large number of sensors such as an air flow meter and a large number of maps using output values from these sensors as arguments. However, when the in-cylinder charged air amount Mc is estimated using the map in this way, in order to make the estimated value of the in-cylinder charged air amount Mc more accurate, the number of necessary maps and their arguments are required. The number of will increase. If the number of maps increases in this way, the ROM of the ECU for storing the maps must have a large storage capacity, which increases the manufacturing cost of the control device for the internal combustion engine. Furthermore, in order to create each map, a calibration operation must be performed for each type of internal combustion engine in which the map is used, but the number of measurement points in this calibration operation increases according to the number of maps and the number of arguments thereof. If the number of maps and the number of arguments increase, the number of man-hours for fitting work will increase.

  In view of this, a control device for an internal combustion engine that calculates the in-cylinder charged air amount Mc by numerical calculation using various models without using a map has been studied. In such a control apparatus, the number of necessary maps is reduced as much as possible by using a lot of numerical calculations. This greatly reduces the number of man-hours for performing the fitting work, but also the in-cylinder charged air amount Mc Can be calculated accurately. Among such control devices, one proposed by the applicant of the present application is a control device equipped with the air model shown in FIG. The illustrated air model M1 is the simplest model applied to an internal combustion engine, and the air model M1 will be described below.

  As shown in FIG. 5, the air model M1 includes a throttle model M10, an intake pipe model M20, and an intake valve model M30. The throttle model M10 includes an opening (throttle valve opening) θt of the throttle valve 19 detected by the throttle valve opening sensor 44 and an atmospheric pressure (or intake pipe) around the internal combustion engine detected by the atmospheric pressure sensor 43. 16) Pa, the atmospheric temperature around the internal combustion engine detected by the intake air temperature sensor 42 (or the temperature of air sucked into the intake pipe 16) Ta, and an intake pipe model M20 described later. The calculated pressure in the intake pipe portion 23 (hereinafter referred to as “intake pipe pressure Pm”) is input, and the values of these input parameters are substituted into a model calculation formula of a throttle model M10 described later. The flow rate of air passing through the throttle valve 19 per unit time (hereinafter referred to as “throttle passage air flow rate mt”) is calculated. The throttle passage air flow rate mt calculated in the throttle model M10 is input to the intake pipe model M20.

  The intake pipe model M20 includes a throttle passage air flow rate mt calculated in the throttle model M10 and a flow rate of air flowing into the combustion chamber 5 per unit time described in detail below (hereinafter referred to as “cylinder intake air flow rate mc”. And the values of the input parameters are substituted into a model calculation formula of an intake pipe model M20, which will be described later, so that the pressure of the air existing in the intake pipe portion 23 (hereinafter, “ And the temperature of the air existing in the intake pipe portion 23 (hereinafter referred to as “intake pipe temperature Tm”). The intake pipe pressure Pm and the intake pipe temperature Tm calculated in the intake pipe model M20 are both input to the intake valve model M30, and the intake pipe pressure Pm is also input to the throttle model M10.

  In addition to the intake pipe pressure Pm and the intake pipe temperature Tm calculated in the intake pipe model M20, an atmospheric temperature Ta is input to the intake valve model M30, and the values of these input parameters are set in the intake valve model M30 described later. By substituting in the model calculation formula, the cylinder intake air flow rate mc is calculated. The calculated in-cylinder intake air flow rate mc is converted into the in-cylinder charged air amount Mc, and the fuel injection amount from the fuel injection valve is determined based on the in-cylinder charged air amount Mc. The in-cylinder intake air flow rate mc calculated in the intake valve model M30 is input to the intake pipe model M20.

  As can be seen from FIG. 5, in the air model M1, the parameter value calculated in one model is used as an input value to another model. Therefore, in the entire air model M1, the actually input value is the throttle valve opening. There are only three parameters of θt, atmospheric pressure Pa, and atmospheric temperature Ta, and the in-cylinder charged air amount Mc is calculated from these three parameters.

Next, each model M10 to M30 of the air model M1 will be described.
In the throttle model M10, the throttle passage air flow rate mt is calculated from the atmospheric pressure Pa, the atmospheric temperature Ta, the intake pipe pressure Pm, and the throttle valve opening degree θt based on the following equation (1). Here, μt in the equation (1) is a flow coefficient in the throttle valve, which is a function of the throttle valve opening θt, and is thus determined from a map as shown in FIG. Further, At indicates the opening cross-sectional area of the throttle valve, which is a function of the throttle valve opening θt, and is determined from a map as shown in FIG. R is a constant related to the gas constant, and is actually a value obtained by dividing the gas constant by the mass Mlmol of gas (air) per mol.

Φ (Pm / Pa) is a function shown in the following formula (2), and κ in the formula (2) is a specific heat ratio (a constant value). Since this function Φ (Pm / Pa) can be expressed in a graph as shown in FIG. 8, such a graph is stored as a map in the ROM 34 of the ECU 31 and is actually calculated using the equation (2). Alternatively, the value of Φ (Pm / Pa) may be obtained from the map.

  Expressions (1) and (2) of the throttle model M10 are such that the gas pressure upstream of the throttle valve 19 is the atmospheric pressure Pa, the gas temperature upstream of the throttle valve 19 is the atmospheric temperature Ta, and the gas downstream of the throttle valve 19 is Applying the law of conservation of mass, the law of conservation of energy and the law of conservation of momentum to the model of the throttle valve 19 as shown in FIG. , And by using the Mayer relation.

In the intake pipe model M20, the intake pipe internal pressure Pm and the intake pipe internal temperature Tm are calculated from the throttle passage air flow rate mt, the cylinder intake air flow rate mc, and the atmospheric temperature Ta based on the following formulas (3) and (4). The Vm in the equations (3) and (4) is a constant equal to the volume of the intake branch pipe 14, the surge tank 15, the intake pipe 16 and the like (intake pipe portion 23) from the throttle valve 19 to the intake valve 6. is there.

Here, the intake pipe model M20 will be described with reference to FIG. When the total gas amount (total air amount) in the intake pipe portion 23 is M, the temporal change in the total gas amount M is the flow rate of the gas flowing into the intake pipe portion 23, that is, the throttle passage air flow rate mt, and the intake pipe portion. 23 is equal to the difference between the flow rate of the gas flowing out from the cylinder 23, that is, the in-cylinder intake air flow rate mc, and the following equation (5) is obtained by the law of conservation of mass. From (Pm · Vm = M · R · Tm), Equation (3) is obtained.

Further, the temporal change amount of the gas energy M · Cv · Tm in the intake pipe portion 23 is equal to the difference between the energy of the gas flowing into the intake pipe portion 23 and the energy of the gas flowing out of the intake pipe portion 23. Therefore, when the temperature of the gas flowing into the intake pipe portion 23 is the atmospheric temperature Ta and the temperature of the gas flowing out from the intake pipe portion 23 is the intake pipe temperature Tm, the following equation (6) is obtained from the energy conservation law. Equation (4) is obtained from Equation (6) and the gas equation of state.

In the intake valve model M30, the in-cylinder intake air flow rate mc is calculated from the intake pipe internal pressure Pm, the intake pipe internal temperature Tm, and the atmospheric temperature Ta based on the following equation (7). In the equation (7), a and b are intake valves in the case of an internal combustion engine having a variable valve mechanism that can change the phase angle (valve timing) and operating angle of the intake valve 6 from the engine speed Ne. 6 is a value determined from the phase angle and the working angle.

  The above-described intake valve model M30 will be described with reference to FIG. In general, the in-cylinder charged air amount Mc, which is the amount of air charged in the combustion chamber 5 when the intake valve 6 is closed, is determined when the intake valve 6 is closed (when the intake valve is closed). This is proportional to the pressure in the combustion chamber 5 when the intake valve 6 is closed. Further, the pressure in the combustion chamber 5 when the intake valve 6 is closed can be regarded as being equal to the pressure of the gas upstream of the intake valve 6, that is, the intake pipe pressure Pm. Therefore, the cylinder charge air amount Mc can be approximated as being proportional to the intake pipe pressure Pm.

  Here, the average air flow rate flowing out of the intake pipe portion 23 per a certain time (for example, a crank angle of 720 °), or the intake pipe portion 23 per a certain time (for example, a crank angle of 720 °). If the cylinder intake air flow rate mc (described in detail below) is obtained by dividing the amount of air sucked into the combustion chambers 5 of all cylinders by the predetermined time, the cylinder charge air amount Mc becomes the intake pipe pressure Pm. Therefore, it is considered that the in-cylinder intake air flow rate mc is also proportional to the intake pipe pressure Pm. From this, the above formula (7) is obtained based on the theory and empirical rules. The value a in the equation (7) is a proportional coefficient, and the value b is a value representing the burned gas remaining in the combustion chamber 5 (the amount of burned gas remaining in the combustion chamber 5 when the exhaust valve 8 is closed). Is equivalent to a value divided by time ΔT180 ° described later). In actual operation, the intake pipe temperature Tm may change greatly during transient operation, and Ta / Tm derived based on theory and empirical rule is multiplied as a correction for this.

  Here, the cylinder intake air flow rate mc will be described with reference to FIG. 12 when the internal combustion engine has four cylinders. In FIG. 12, the horizontal axis represents the rotation angle of the crankshaft, and the vertical axis represents the flow rate of air actually flowing from the intake pipe portion 23 into the combustion chamber 5 per unit time. As shown in FIG. 12, in the four-cylinder internal combustion engine, the intake valve 6 is opened in the order of, for example, the first cylinder, the third cylinder, the fourth cylinder, and the second cylinder, and the intake valve 6 corresponding to each cylinder is opened. Air flows from the intake pipe portion 23 into the combustion chamber 5 of each cylinder according to the valve opening amount. For example, the displacement of the flow rate of air flowing into the combustion chamber 5 of each cylinder from the intake pipe portion 23 is as shown by a broken line in FIG. The flow rate of the air flowing into is as shown by the solid line in FIG. Further, for example, the in-cylinder charged air amount Mc to the first cylinder is as shown by hatching in FIG.

  On the other hand, the in-cylinder intake air flow rate mc is obtained by averaging the flow rate of the air flowing into the combustion chambers 5 of all the cylinders from the intake pipe portion 23 shown by the solid line, and is indicated by a one-dot chain line in the drawing. In the cylinder intake air flow rate mc indicated by the one-dot chain line, in the case of four cylinders, the crankshaft is 180 ° (that is, the angle 720 ° at which the crankshaft rotates during one cycle in the four-stroke internal combustion engine) The in-cylinder charged air amount Mc is obtained by multiplying the time ΔT180 ° required for rotation by an angle divided by the number. Therefore, the cylinder intake air amount Mc is calculated by multiplying the cylinder intake air flow rate mc calculated by the intake valve model M30 by ΔT180 ° (Mc = mc · ΔT180 °). More specifically, in consideration of the fact that the in-cylinder charged air amount Mc is proportional to the pressure when the intake valve is closed, the in-cylinder intake air flow rate mc when the intake valve is closed is multiplied by ΔT180 °. The filling air amount Mc.

Next, a case where the air model M1 is mounted on a control device for an internal combustion engine and the in-cylinder charged air amount Mc is actually calculated will be described. The in-cylinder charged air amount Mc is expressed by solving the above formula (1), formula (3), formula (4), and formula (7) using the air model M1. In this case, in order to be processed by the ECU 31, these equations need to be discretized. When the formula (1), the formula (3), the formula (4), and the formula (7) are discretized using the time t and the calculation interval Δt, the following formulas (8), (9), and (10) are respectively obtained. And Equation (11) is obtained. The intake pipe internal temperature Tm (t + Δt) is calculated by equation (12) from Pm / Tm (t + Δt) and Pm (t + Δt) calculated by equations (9) and (10), respectively.

  In the air model M1 implemented in this way, the throttle passage air flow rate mt (t) at time t calculated by the equation (8) of the throttle model M10 and the time calculated by the equation (11) of the intake valve model M30. The in-cylinder intake air flow rate mc (t) at t is substituted into the equations (9) and (10) of the intake pipe model M20, whereby the intake pipe internal pressure Pm (t + Δt) and the intake pipe internal temperature Tm (at time t + Δt) t + Δt) is calculated. Next, the calculated Pm (t + Δt) and Tm (t + Δt) are substituted into the equations (8) and (11) of the throttle model M10 and the intake valve model M30, whereby the throttle passage air flow rate mt (t + Δt) at time t + Δt. ) And the cylinder intake air flow rate mc (t + Δt). Then, by repeating such calculation, the cylinder intake air flow rate mc at any time t is calculated from the throttle valve opening θt, the atmospheric pressure Pa, and the atmospheric temperature Ta, and the calculated cylinder intake air flow rate is calculated. By multiplying mc by the time ΔT180 °, the in-cylinder charged air amount Mc at an arbitrary time t is calculated.

  At the time of starting the internal combustion engine, that is, at time t = 0, the intake pipe pressure Pm is equal to the atmospheric pressure (Pm (0) = Pa), and the intake pipe temperature Tm is equal to the atmospheric temperature (Tm (0)). = Ta), the calculation in each of the models M10 to M30 is started.

  In the air model M1, the atmospheric temperature Ta and the atmospheric pressure Pa are assumed to be constant. However, the air model M1 may be a value that changes depending on the time, for example, detected at the time t by an intake temperature sensor for detecting the atmospheric temperature. Assuming that the value is the atmospheric temperature Ta (t) and the value detected at time t by the atmospheric pressure sensor for detecting the atmospheric pressure is the atmospheric pressure Pa (t), the above equations (8), (10), and (11) ) May be substituted.

  As described above, the air model M1 is the simplest model applied to an internal combustion engine, and the cylinder charge air amount Mc calculated by the air model M1 has an error with respect to the actual cylinder charge air amount. easy. Therefore, in the air model M2 shown in FIG. 13 to be described later, in addition to the above three parameters of the throttle valve opening θt, the atmospheric pressure Pa, and the atmospheric temperature Ta, the air model calculates the in-cylinder charged air amount with higher accuracy. Thus, the in-cylinder charged air amount Mc is calculated based on the output value AFM of the air flow meter 41. Hereinafter, the air model M2 will be described.

  The air model M2 includes an electronically controlled throttle model M40 and an air flow meter model (AFM model) M50 in addition to the throttle model M10, the intake pipe model M20, and the intake valve model M30. The electronically controlled throttle model M40 receives the output value Accp of the depression amount sensor 46 that generates an output voltage proportional to the depression amount of the accelerator pedal 45, and based on this, the throttle valve at a time T0 ahead of the current time. A throttle valve opening (hereinafter referred to as “throttle valve opening after a predetermined time”) θtf that is expected to reach 19 is calculated.

  Further, the AFM model M50 receives a throttle passage air flow rate (hereinafter referred to as an “estimated value of the throttle passage air flow rate”) mt calculated by the throttle model M10, and this input value is used as an AFM model M50 described later. By substituting this into the model calculation formula, the air flow meter 41 outputs when it is assumed that air is actually flowing through the intake pipe 16 where the throttle valve 19 is disposed for the estimated value mt of the throttle passage air flow rate. An output value (hereinafter referred to as “expected output value”) AFMmt that will be calculated. In the following description, a value calculated by each model M10 or the like is referred to as an “estimated value”.

  That is, the output of the air flow meter 41 is specific to the actual air flow passing air flow rate (the air flow passing air flow rate and the throttle passing air flow rate are considered to be substantially the same, and will be described below as the throttle passing air flow rate). There is a response delay based on the response characteristics, and therefore, the value does not correspond only to the current actual throttle passage air flow rate. The AFM model M50 is a model simulating this response characteristic, and calculates an expected output value in consideration of the response delay of the air flow meter 41.

  Here, the air model M2 can be divided into three model blocks M2 ′, M2 ″, and M2 ′ ″ as shown in FIG. In the model block M2 ′, an estimated value θtf of the throttle valve opening after a predetermined time is calculated by the electronic control throttle model M40 based on the output value Accp of the depression amount sensor 46, and the throttle valve after the calculated predetermined time Based on the estimated value θtf of the opening, an estimated value mtf of the throttle passage air flow after a predetermined time is calculated by the throttle model M10 ′. Then, the estimated value Pmf of the intake pipe pressure after the predetermined time is calculated by the intake valve model M20 ′ based on the calculated estimated value mtf of the throttle passage air flow after the predetermined time. The calculated estimated value Pmf of the intake pipe pressure after a predetermined time is input to the throttle model M10 ′ and the intake valve model M30 ′. In the intake valve model M30 ′, an estimated value mcf of the in-cylinder intake air flow after a predetermined time is calculated based on the estimated value Pmf of the intake pipe pressure after a predetermined time, and is input to the intake pipe model M20 ′.

  Therefore, in the entire model block M2 ′ of the air model M2, by inputting the throttle valve opening θtf after a predetermined time, the estimated value mcf of the cylinder intake air flow after the predetermined time and the estimation of the intake pipe pressure after the predetermined time are estimated. A value Pmf is calculated. Thus, by calculating the estimated value mcf of the in-cylinder intake air flow after a predetermined time, it becomes easy to make the actual air-fuel ratio coincide with the target air-fuel ratio. That is, during the transient operation of the internal combustion engine, the in-cylinder intake air flow rate changes every moment, so that the current in-cylinder intake air flow rate (or the in-cylinder charged air amount calculated from the current in-cylinder intake air flow rate) is obtained. Even if the fuel injection amount is calculated based on this, the cylinder intake air flow rate may have already changed when the fuel is actually injected, and even if the current cylinder intake air flow rate can be accurately calculated, the result Therefore, the actual air / fuel ratio may differ from the target air / fuel ratio. On the other hand, if the fuel injection amount is calculated based on the in-cylinder intake air flow rate after a predetermined time, the in-cylinder intake air flow rate is obtained when the fuel is actually injected. It becomes easy to match the fuel ratio.

  However, since the model block M2 ′ performs the same calculation as the air model M1 except that the electronic control throttle model M40 exists, the in-cylinder intake air flow rate after a predetermined time calculated in the model block M2 ′ When the in-cylinder charged air amount Mc is calculated based on the estimated value mcf, an error is likely to occur with respect to the actual in-cylinder charged air amount as described above.

  Therefore, in the air model M2, the estimated value Pmf of the intake pipe pressure calculated in the model block M2 ′ is corrected based on the output of the air flow meter 41, and the in-cylinder charged air amount Mc is calculated based on the corrected intake pipe pressure Pmfin. We are going to calculate.

  That is, in the model block M2 ″, the same calculation as that of the air model M1 is performed based on the current throttle valve opening θtc detected by the throttle valve opening sensor 44. That is, the current throttle valve opening θtc is input to the throttle model M10 ″ and the estimated value mtc of the current throttle passage air flow rate is output, and the intake pipe model M20 ″ has the current throttle passage air flow rate. An estimated value mtc is input and an estimated value Pmc of the current intake pipe pressure is output, and the intake pipe pressure Pmc is input to the throttle model M10 ″ and the intake valve model M30 ″. In the intake valve model M30 ″, an estimated value mcc of the current in-cylinder intake air flow rate is calculated and input to the intake pipe model M20 ″.

  Based on the estimated value mtc of the current throttle passage air flow calculated in the above calculation, the AFM model M50 calculates the expected output value AFMmt that takes into account the response delay of the air flow meter 41, and thus the predicted value thus calculated. The output value AFMmt is input to the intake pipe model M20 ′ ″ as the throttle passage air flow rate. Then, the estimated value Pmmdl of the intake pipe pressure is calculated by the intake pipe model M20 ′ ″ and the intake valve model M30 ′ ″ by the same method as described above. The estimated value Pmmdl of the intake pipe pressure calculated in this way is output by the air flow meter 41 calculated using the same air model as the model block M2 ′ based on the current throttle valve opening θtc. It shows the intake pipe pressure when it is assumed that the flow rate of air passing through the throttle is equal to the wax output value (hereinafter referred to as “estimated value of intake pipe pressure using the AFM model”).

  On the other hand, in the model block M2 ′ ″, the actual output value AFM of the air flow meter 41 is inputted to the intake pipe model M20 ″ ″ as the throttle passage air flow rate, and the intake pipe model M20 ″ ″ controls the intake pipe pressure. An estimated value Pmafm is output. The estimated value Pmafm of the intake pipe pressure is input to the intake valve model M30 ″ ″, the in-cylinder intake air flow rate mcafm is output, and is input again to the intake pipe model M20 ″ ″. The estimated value Pmafm of the intake pipe pressure calculated in this way indicates the intake pipe pressure when the throttle passage air flow rate is assumed to be equal to the output value of the air flow meter 41 (hereinafter referred to as “the output of the air flow meter”). Based on the estimated value of the intake pipe pressure).

  Therefore, the difference between the estimated value Pmmdl of the intake pipe pressure using the AFM model and the estimated value Pmafm of the intake pipe pressure based on the output of the air flow meter is based on the current intake pipe pressure calculated by the air model and the air flow meter 41. It represents an error from the calculated intake pipe pressure. Therefore, the estimated value Pmfin (= Pmf + Pmafm−Pmmdl) of the intake pipe pressure obtained by adding this difference to the estimated value Pmf of the intake pipe pressure calculated in the model block M2 ′ substantially matches the actual intake pipe pressure. The estimated value Pmfin of the intake pipe pressure is input to the intake pipe model M30, and the estimated value Mc of the in-cylinder charged air amount is calculated based on the estimated value mcfin of the in-cylinder intake air flow rate calculated by the intake pipe model M30. Then, the fuel injection amount is calculated based on the estimated value Mc of the in-cylinder charged air amount.

Next, the electronic control throttle model M40 and the AFM model M50 will be described.
The electronically controlled throttle model M40 is a model that calculates the throttle valve opening degree θtf after a predetermined time T0 from the current time based on the depression amount of the accelerator pedal 45. In the present embodiment, the provisional target throttle valve opening θr1 is obtained based on the accelerator pedal depression amount Accp detected by the depression amount sensor 46 and the map shown in FIG. A value obtained by delaying the valve opening θr1 by a predetermined time T (for example, 64 msec) is determined as the final target throttle valve opening θr. Then, the ECU 31 sends a drive signal to the step motor 18 so that the actual throttle valve opening becomes the target throttle valve opening θr.

  Thus, the target throttle valve opening degree θr at the current time is equal to the provisional target throttle valve opening degree θr1 determined according to the accelerator pedal depression amount Accp at the time before the predetermined time T from the current time. The target throttle valve opening θr at a time after a predetermined time T0 (0 ≦ T0 ≦ T) from the current time is equal to the temporary target throttle valve opening θr1 before the time (T−T0) from the current time. If the operation delay time of the step motor 18 is ignored, the target throttle valve opening θr is equal to the actual throttle valve opening. Based on this idea, in the electronically controlled throttle model M40, the provisional target throttle valve opening θr1 before the time (T-T0) from the current time is set to the throttle valve opening at the time t after the predetermined time T0 from the current time. Degree (throttle valve opening after a predetermined time) θtf. In addition to the operation delay time of the step motor 18, the throttle valve opening θtf after a predetermined time may be set.

  Next, the AFM model M50 will be specifically described. In the air flow meter 41, as described above, the power supplied to the heating resistor 41a3 is adjusted so that the temperature difference between the intake temperature measuring resistor 41a1 and the heating resistor (bobbin portion) 41a3 is always kept constant. The airflow passage air flow rate is calculated based on the supplied power. Here, since the supplied power indicates the amount of heat released from the bobbin portion 41a3 and the support portion 41a4 to the air passing through the intake pipe 16, the air flow meter 41 is based on the amount of heat released from the bobbin portion 41a3 and the support portion 41a4. In other words, the airflow passing air flow rate is calculated.

  More specifically, the bobbin portion 41a3 is formed by winding a platinum hot wire around a cylindrical ceramic bobbin and coating the outer periphery thereof with glass. For this reason, heat dissipation from the platinum heat wire to the surrounding air is delayed due to the glass layer interposed between the platinum heat wire and the surrounding air. Therefore, even when the flow rate of the air passing through the intake pipe 16 suddenly increases, the amount of heat released from the bobbin portion 41a3 per unit time (hereinafter simply referred to as “heat dissipation amount”) is not immediately available. It does not increase, but increases with some delay. In other words, the amount of heat released from the bobbin portion 41a3 has a response delay with respect to the actual airflow passing air flow rate. The same applies to the amount of heat released from the support portion 41a4. The amount of heat released from the support portion 41a4 has a response delay with respect to the actual airflow passing air flow rate.

It is known that this response delay can be approximated to a first-order delay, and the response delay of the heat radiation amount of the bobbin portion 41a3 is expressed by the following equation (13). Here, ω _b in the equation (13) is a heat dissipation amount of the bobbin portion 41a3 converted from the output value of the air flow meter 41, that is, a heat dissipation amount actually radiated from the platinum heat wire of the bobbin portion 41a3 as a result of the heat dissipation delay (hereinafter referred to as “heat dissipation amount”). , Referred to as “delayed heat release amount”). W _b in the equation (13) is a heat dissipation amount compensated for the response delay, that is, a heat dissipation amount radiated from the platinum heat wire of the bobbin portion 41a3 when it is assumed that no heat dissipation delay occurs (hereinafter referred to as “complete heat dissipation amount”). Designated). That is, the complete heat radiation amount W _b is equal to the heat radiation amount when the internal combustion engine is in steady operation, and is basically a function of only the flow rate of air passing through the intake pipe 16 near the air flow meter 41. Since the flow rate of air passing through the intake pipe 16 near the air flow meter 41 is substantially equal to the flow rate of air passing through the throttle, the complete heat radiation amount W _b can be considered as a function of the flow rate of air passing through the throttle. relationship between the true heat radiation amount W _b from can be represented as in Figure 15. Further, τ _b in the equation (13) is a time constant of a first-order lag in heat radiation from the bobbin portion 41a3, and a calculation method thereof will be described later.

Similarly, the response delay of the heat radiation amount of the support portion 41a4 is expressed by the following equation (14). The complete heat release amount W _b from the support portion 41a4 can also be considered as a function of the throttle passage air flow rate, and the relationship between the throttle passage air flow rate and the complete heat release amount W _s from the support portion 41a4 can be expressed as shown in FIG. it can. Further, τ _s in the equation (14) is a time constant of a first-order lag in heat dissipation from the support portion 41a4.

Delayed heat radiation amounts ω _b and ω _s from the bobbin portion 41a3 and the support portion 41a4 are calculated by the equations (13) and (14). In the air flow meter 41, the output voltage changes according to the amount of heat radiated from the bobbin portion 41a3 and the support portion 41a4 as a result of the heat radiation delay, that is, the amount of delayed heat radiation, so the equations (13) and (14) The output value (expected output value) AFMmt that the airflow meter 41 will output is calculated based on the sum of the delayed heat radiation amount calculated by (ω _b + ω _s ). In the present embodiment, the relationship between the sum of the delayed heat release amount (ω _b + ω _s ) in the air flow meter 41 and the expected output value AFMmt of the air flow meter 41 is obtained in advance experimentally or by calculation, and a map as shown in FIG. Is stored in the ROM 34 of the ECU 31. Then, the expected output value AFMmt of the air flow meter 41 is calculated using the map based on the sum of delayed heat dissipation ω _b + ω _s calculated from the equations (13) and (14). The calculation of the expected output value AFMmt of the air flow meter 41 based on the sum of delayed heat dissipation ω _b + ω _s using the above map can be expressed as the following equation (15).

The primary delay time constant τ _b of the bobbin portion 41a3 is calculated by the following equation (16), and the primary delay time constant τ _s of the support portion 41a4 is calculated by the following equation (17). In Expression (16) and Expression (17), u is an air flow rate per unit cross-sectional area in the flow path in the detection unit of the air flow meter 41, that is, in the bypass passage of the air flow meter 41. The air flow rate u per unit sectional area is calculated based on the output value AFM of the air flow meter 41 by a map or by a predetermined calculation formula. K _b , k _s , m _b , and m _s are constants obtained in advance by experiment or calculation, k _b and m _b are constants for the bobbin portion 41a3, and k _s and m _s are constants for the support portion 41a4. Respectively. Since the degree of response delay differs between the bobbin portion 41a3 and the support portion 41a4, a calculation for calculating an expected output value from the air flow rate through the throttle by setting the time constant by separating the bobbin portion 41a3 and the support portion 41a4. The accuracy is to be improved.

Next, a description will be given of a case where the AFM model M50 is mounted on a control device for an internal combustion engine and the expected output value AFMmt of the air flow meter 41 is calculated from the estimated value mt of the air flow rate through the throttle actually calculated by the throttle model M10. To do. The expected output value AFMmt of the air flow meter 41 is expressed by solving the above equations (13) and (14) using the AFM model M50. In this case, in order to be processed by the ECU 31, these equations need to be discretized. When the equations (13) and (14) are discretized using the time t and the calculation interval Δt, the following equations (18) and (19) are obtained, respectively.

In the AFM model M50 mounted in this way, the delayed heat release amount ω _b (t) from the bobbin portion 41a3, the delayed heat release amount ω _s (t) from the support portion 41a4 at time t, and the formula (8 ), The estimated value mt (t) of the air flow rate through the throttle at time t calculated in step (8) is substituted into the equations (18) and (19), whereby the delayed heat release ω _b (t + Δt from the bobbin portion 41a3 at time t + Δt. ) And the delayed heat radiation amount ω _s (t + Δt) from the support portion 41a4. Then, the predicted output value AFMmt (t + Δt) of the air flow meter 41 at time t + Δt is calculated by the equation (15) using these delayed heat radiation amounts ω _b (t + Δt) and ω _s (t + Δt).

  By the way, in the apparatus shown in FIGS. 5 and 13 as described above, in-cylinder filling is performed by sequential calculation from the start of the engine using a model calculation formula for inputting the throttle valve opening θt, the atmospheric temperature Ta, the atmospheric pressure Pa, and the like. An intake air amount is calculated. For this reason, if there is a modeling error in each element model of the air model, the calculated in-cylinder charged intake air amount becomes inaccurate and the fuel injection amount cannot be accurately calculated.

  On the other hand, considering each element model of the air model (throttle model M10, intake pipe model M20, intake valve model M30, electronically controlled throttle model M40, AFM model M50), even if there is no error in the modeling itself, Modeling errors may occur due to changes in characteristics due to use or variations due to manufacturing tolerances. Therefore, in this embodiment, each element model is corrected. In particular, in this embodiment, the throttle model M10 and the AFM model M50 are corrected.

  Hereinafter, the model correction method of this embodiment will be described. In the present embodiment, the current actual airflow passing air flow rate, throttle valve opening, and intake pipe pressure are measured by the airflow meter 41, throttle valve opening sensor 44, and intake pressure sensor 40, respectively, and these values are used. A correction operation is performed on the throttle model M10, the AFM model M50, and the like independently of the cylinder intake air amount calculation in FIG.

  First, the correction operation of the throttle model M10 will be described. FIG. 17 is a block diagram illustrating a correction operation of the throttle model M10. As shown in FIG. 17, the AFM model M50 is used when the throttle model is corrected.

  In the present embodiment, the throttle model M10 (with the current intake pipe pressure Pm and the throttle valve opening θt (and the atmospheric pressure Pa, the atmospheric temperature Ta) measured by the intake pressure sensor 40 and the throttle valve opening sensor 44 are used. The formula (1) of the throttle model M10 is corrected so that the estimated value mt of the throttle passage air flow rate calculated by FIG. 5) matches the current throttle passage air flow rate calculated from the output of the air flow meter 41.

  That is, in FIG. 17, when the estimated value mt of the throttle passage air flow rate calculated by the equation (1) of the throttle model M10 using the actual measurement value Pm and the throttle valve opening θt is input to the AFM model M50, the actual throttle It is converted into an expected output value AFMmt of the air flow meter 41 when it is assumed that the passing air flow rate is the estimated value mt of the throttle passing air flow rate.

  Therefore, when the modeling error is not included in the throttle model M10, that is, when the actual throttle passing air flow rate matches the estimated value mt of the throttle passing air flow rate calculated by the throttle model M10, The output value AFM of the air flow meter 41 should match the expected output value AFMmt of the AFM model M50. If they do not match, the actual throttle passage air flow rate and the estimated value mt of the throttle passage air flow rate are equal. Therefore, it is considered that a modeling error has occurred in the throttle model M10.

  Therefore, when the output value AFM and the predicted output value AFMmt do not match, the predicted output value AFMmt is the output value (actual value) AFM of the air flow meter 41 in order to compensate for the modeling error occurring in the throttle model M10. It is necessary to correct the model calculation formula of the throttle model M10 so that

上記式（２０）のパラメータのうち流量係数μｔは製品毎のばらつきや使用による特性変化の影響を受けやすい。したがって、本実施形態では、スロットルモデルＭ１０の式（２０）により算出されたスロットル通過空気流量の推定値ｍｔをＡＦＭモデルＭ５０の式（１３）および式（１４）に代入して算出された予想出力値ＡＦＭｍｔがエアフロメータ４１の出力値ＡＦＭに一致するように流量係数μｔの値を修正することとしている。これにより、スロットルモデルＭ１０が修正され、スロットルモデルＭ１０を用いた計算結果をエアフロメータ４１の出力値と一致させることができるようになり、よってスロットルモデルＭ１０によって算出されたスロットル通過空気流量の推定値ｍｔを実際のスロットル通過空気流量に一致させることができるようになる。なお、当然ながらスロットルモデルＭ１０の上記修正結果は図５筒内充填吸気量Ｍｃの計算にも同時に反映され、筒内充填吸気量Ｍｃの推定精度が向上するようになる。 Here, in the present embodiment, as described above, the following formula (20) is used as the calculation formula of the throttle model M10.

Among the parameters of the above equation (20), the flow coefficient μt is easily affected by variations among products and characteristic changes due to use. Therefore, in this embodiment, the expected output calculated by substituting the estimated value mt of the throttle passage air flow rate calculated by the equation (20) of the throttle model M10 into the equations (13) and (14) of the AFM model M50. The value of the flow coefficient μt is corrected so that the value AFMmt matches the output value AFM of the air flow meter 41. As a result, the throttle model M10 is corrected, and the calculation result using the throttle model M10 can be made to coincide with the output value of the air flow meter 41. Therefore, the estimated value of the throttle passage air flow rate calculated by the throttle model M10. It becomes possible to make mt equal to the actual throttle passage air flow rate. Of course, the correction result of the throttle model M10 is also reflected in the calculation of the in-cylinder charged intake air amount Mc in FIG. 5, so that the estimation accuracy of the in-cylinder charged intake air amount Mc is improved.

As a method for correcting the value of the flow coefficient μt, for example, the following equation (21) is conceivable. Note that μt ′ in the equation (21) is a corrected flow coefficient.

  However, if the value of the flow coefficient μt is corrected by the equation (21), it can be appropriately corrected when the internal combustion engine is in steady operation, but it should be appropriately corrected when the internal combustion engine is in transient operation. I can't. Hereinafter, this will be briefly described with reference to FIG.

FIG. 18 is a time chart of the throttle valve opening θt and the values of various parameters relating to the air flow rate. In the figure, time t ₀ shows when the throttle valve opening θt becomes θt _0, and time t ₁ shows when the throttle valve opening θt becomes θt ₁ . 18 shows that the actual throttle passage air flow rate (solid line in FIG. 18) and the estimated value mt of throttle passage air flow rate calculated by the throttle model M10 (dotted line in FIG. 18) are before time t _0. The time chart is shown in the case where the values are substantially the same after time t ₀ . That is, FIG. 18 shows an estimated value of the throttle passage air flow rate calculated by the throttle model M10 because an error has occurred in the value of the flow coefficient μt before the time t ₀ , that is, when the throttle valve opening θt is smaller than θt _0. When mt has deviated from the actual air flow rate through the throttle and after time t ₀ , that is, when the throttle valve opening θt is greater than or equal to θt ₀ , there is almost no error in the value of the flow coefficient μt. The time chart in the case where the estimated value mt of the calculated throttle passage air flow rate substantially matches the actual throttle passage air flow rate is shown.

  As described above, since the air flow meter 41 has a response delay due to a heat release delay, when the actual flow rate of air passing through the throttle changes as shown in FIG. 18 is a delay with respect to the actual throttle passage air flow rate. In particular, considering that the output value AFM of the air flow meter 41 can be approximated with a first-order lag with respect to the actual throttle passage air flow rate, the current output value AFM of the air flow meter 41 is only the current actual throttle passage air flow rate. It does not change according to the above, but also changes according to the past actual throttle passage air flow history.

  Further, as described above, in the AFM model M50, the response delay based on the heat release delay of the air flow meter 41 is taken into consideration, and therefore, when the actual flow rate of air passing through the throttle changes as shown in FIG. The expected output value AFMmt calculated by the model M50 is also delayed with respect to the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10.

For this reason, as shown in FIG. 18, the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10 after time t ₀ substantially coincides with the actual throttle passage air flow rate. After t ₀ , the output value AFM of the air flow meter 41 does not match the expected output value AFMmt calculated by the AFM model M50. Therefore, for example, if the flow coefficient μt of the throttle model M10 is to be corrected by the equation (21) using the output value AFM and the predicted output value AFMmt, the flow coefficient μt has no error after time t _0. Therefore, the flow coefficient μt is corrected.

Alternatively, as a method for correcting the value of the flow coefficient μt, for example, the following equation (22) is conceivable. Thus, by using the integration, it is possible to suppress the influence of noise existing in the output value AFM of the air flow meter 41.

  However, even when the value of the flow coefficient μt is corrected by the equation (22), it can be appropriately corrected when the internal combustion engine is in steady operation, but when the internal combustion engine is in transient operation. It cannot be corrected properly. This will be briefly described below.

  Here, although the map of the flow coefficient μt is shown continuously in FIG. 6, it is actually divided into certain engine operating areas, and the map value of the flow coefficient μt is set for each engine operating area. The Specifically, the map of the flow coefficient μt is obtained by dividing the throttle valve opening θt into a plurality of map areas as shown in FIG. 19, and one map of the flow coefficient μt for each map area of the throttle valve opening θt. Value is set. Therefore, when the flow coefficient μt is corrected, the value of the flow coefficient μt corresponding to the map area is corrected for each map area of the throttle valve opening θt.

Referring to the example shown in FIG. 18, when one map region of the throttle valve opening θt is θt ₀ to θt ₁ , the period during which the throttle valve opening θt is within the map region is from time t ₀ to time t. It is up to ₁ . Therefore, the error rate of the value of the flow coefficient μt when the throttle valve opening θt is in this map region is determined by the average value of the output value AFM of the air flow meter 41 from the time t ₀ to the time t ₁ and the AFM model M50. When calculated based on the ratio of the calculated expected output value AFMmt to the average value, it is considered that the error in the value of the flow coefficient μt can be compensated relatively accurately. Based on this idea, the equation (22) The correction of the value of the flow coefficient μt by has been proposed. More specifically, in the equation (22), the ratio between the average value of the output value AFM of the air flow meter 41 and the average value of the expected output value AFMmt is the integrated value of the output value AFM of the air flow meter 41 and the expected output value thereof. Since it is equal to the ratio with the integrated value of AFMmt, in equation (22), the integrated value of the output value AFM of the air flow meter 41 (area S ₁ + S in FIG. 18) is added to the flow coefficient μt (θt) currently stored in the map. ₂ ) is multiplied by an integral value of the expected output value AFMmt (area S _{1 in} FIG. 18) (θt · (S ₁ + S ₂ ) / S ₁ ) to correct the flow coefficient μt (θt) The calculated value μt ′ (θt) is calculated.

However, as described above, in the example shown in FIG. 18, the actual throttle passage air flow rate and the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10 substantially coincide after time t ₀ . Nevertheless, since there is a response delay in the air flow meter 41, the integrated value (S ₁ + S ₂ ) of the output value AFM of the air flow meter 41 from time t ₀ to time t ₁ and the expected output calculated by the AFM model M50. The integrated value (S ₁ ) of the value AFMmt from time t ₀ to time t ₁ does not match. For this reason, the flow coefficient μt is corrected even though there is no error in the flow coefficient μt in the map region where the throttle valve opening θt is θt ₁ to θt ₂ .

  As described above, when the value of the flow coefficient μt is corrected by the equations (21) and (22), it can be accurately corrected when the internal combustion engine is in steady operation, but the internal combustion engine is in transient operation. When it is, it cannot be corrected accurately. Therefore, the present invention provides a control apparatus for an internal combustion engine that can accurately correct the value of the flow coefficient μt even when the internal combustion engine is in a transient operation.

FIG. 20 is a time chart of the throttle valve opening θt and the values of various parameters relating to the air flow rate, as in FIG. Hereinafter, with reference to FIG. 20, the correction of the flow coefficient μt in the map region A where the throttle valve opening θt is θt ₀ to θt ₁ will be considered. As shown in the equations (13) and (14), in the AFM model M50, it is assumed that there is a first-order lag with respect to the estimated value mt of the air flow rate through the throttle calculated by the throttle model M10. The value AFMmt is calculated. For this reason, the expected output value AFMmt of the AFM model M50 is calculated based not only on the estimated value mt of the current throttle passage air flow calculated by the throttle model M10 but also on the estimated value mt of the throttle passage air flow before a predetermined time. Will be.

For example, when the throttle valve opening θt at time t ₀ as shown in FIG. 20 has penetrated the map area A, expected output value AFMmt calculated by AFM model M50 at time t ₀ and later, at time t ₀ after Not only the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10 but also the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10 before the time t ₀ is calculated. That is, the expected output value AFMmt calculated by the AFM model M50 is calculated based on the estimated value mt of the throttle passage air flow rate when the throttle valve opening degree θt is in the map region other than the map region A.

Here, as described above, in the example shown in FIG. 20, the flow coefficient μt corresponding to the map area A where the throttle valve opening θt is θt ₀ to θt ₁ is to be corrected, and the flow coefficient μt is corrected. The ratio between the actual throttle passage air flow rate and the estimated throttle passage air flow rate mt calculated by the throttle model M10 is calculated as the ratio between the output value AFM of the air flow meter 41 and the expected output value AFMmt calculated by the AFM model M50. I am trying to calculate from the ratio. However, when the throttle valve opening θt is in the map area A, the expected output value AFMmt calculated by the AFM model M50 is influenced by the estimated value mt of the throttle passage air flow rate in the map area other than the map area A. Therefore, when the throttle valve opening degree θt is in the map area A, the ratio between the output value AFM of the air flow meter 41 and the expected output value AFMmt calculated by the AFM model M50 is that the throttle valve opening degree θt is in the map area A. In this case, the ratio of the actual throttle passage air flow rate to the estimated value mt of the throttle passage air flow rate is not shown.

Therefore, in the present embodiment, when correcting the flow coefficient μt when the throttle valve opening θt is in the map region A, the predicted output value AFMmt calculated by the AFM model M50 is not directly used, but the predicted output value. The influence of the estimated value mt of the air flow rate through the throttle before the throttle valve opening θt enters the map area A from the AFMmt (in the example shown in FIG. 20, when the throttle valve opening θt is smaller than θt ₀ ) is removed. The value (hereinafter referred to as “corrected output value”) AFMmte is used.

In the present embodiment, the calculation of the corrected output value AFMmt is basically performed in the same manner as the calculation of the expected output value AFMmt in the AFM model M50. That is, the delayed heat radiation amounts ω _be and ω _se of the bobbin portion 41a3 and the support portion 41a4 are calculated by the following equations (23) and (24), and the equation (25) is calculated based on the calculated delayed heat radiation amounts ω _be and ω _se. ) To calculate the corrected output value AFMmte from a map similar to the map shown in FIG. Expressions (23) and (24) are discretized as shown in the following expressions (26) and (27) using the time t and the calculation interval Δt for processing by the ECU 31.

However, unlike the calculation of the expected output value AFMmt in the AFM model M50, the correction output value AFMmt is calculated for each map area. That is, when the throttle valve opening θt enters a certain map area, in the example shown in FIG. 20, when the throttle valve opening θt enters the map area A (that is, the throttle valve opening θt becomes θt ₀ ). When the calculation of the corrected output value AFMmte in the map area is started and the throttle valve opening degree θt departs from the map area, the throttle valve opening degree θt is changed from the map area A in the example shown in FIG. The calculation of the corrected output value AFMmte in the map area is completed when the vehicle has left (that is, when the throttle valve opening θt becomes θt ₁ ).

  In calculating the corrected output value AFMmte, an estimated value mt of the throttle passage air flow rate calculated by the throttle model M10 before the throttle valve opening θt enters the map area A (hereinafter referred to as “the past throttle passage air flow rate”). In order to eliminate the influence of the estimated value mt ”, the value of the corrected output value AFMmte (hereinafter referred to as“ initial value ”) when the throttle valve opening θt enters the map area A is referred to as an air flow meter. The output value AFM of 41 is almost the same value.

  That is, as described above, the corrected output value AFMmte is calculated in the ECU 31 based on the discretized equations (26) and (27), but these equations (26) and (27) are other than the initial values. There is no term in which the estimated value mt of the throttle passage air flow rate (before the throttle valve opening θt enters the map area A) is affected. Conversely, as long as the initial value is set to a value that is not affected by the estimated value mt of the past throttle passage air flow rate, the influence of the estimated value mt of the past throttle passage air flow rate on the corrected output value AFMmte is reduced. It can be minimized.

  In the present embodiment, the initial value of the corrected output value AFMmte when the throttle valve opening θt enters the map area A is set to a value almost the same as the output value AFM of the air flow meter 41 at that time, so the corrected output value AFMmte Is not a value determined based on the estimated value mt of the past throttle passage air flow rate, but a value determined based on the output value AFM of the air flow meter 41 when the throttle valve opening θt enters the map area A. Therefore, the influence of the estimated value mt of the past throttle passage air flow rate is removed from the corrected output value AFMmte.

Thus, in order to calculate the subsequent corrected output value AFMmt by the equations (23) and (24) with the initial value of the corrected output value AFMmte being substantially similar to the output value AFM of the air flow meter 41, the time t ₀ it is necessary to match the actual amount of heat dissipated from the actual heat radiation amount and support section 41a4 of the bobbin portion 41a3 at time t ₀ the delay amount of heat radiation delay heat radiation amount and support section 41a4 of the bobbin portion 41a3 in. However, it is difficult to calculate the actual amount of heat radiation from the bobbin portion 41a3 and the support portion 41a4 at time t ₀ from the output value AFM of the air flow meter 41 at time t _0. That is, the total amount of actual heat radiation from the bobbin portion 41a3 and the support portion 41a4 can be obtained from the output value AFM of the air flow meter 41 by using the map shown in FIG. 18 or the like, but the bobbin portion 41a3 and the support portion 41a4 respectively. It is not possible to determine the actual heat dissipation from

  Therefore, in the present embodiment, an observer that can calculate the actual heat radiation amount from each of the bobbin portion 41a3 and the support portion 41a4 of the air flow meter 41 relatively accurately is used. This observer always calculates an approximate value of the actual heat radiation amount from each of the bobbin portion 41a3 and the support portion 41a4 of the air flow meter 41 regardless of the time and the map region of the throttle valve opening θt.

Specifically, the observer uses the following equation (28) and equation (29) to approximate the actual heat radiation amount from the bobbin portion 41a3 (hereinafter referred to as “approximate heat radiation amount”) ω _bo and the support portion 41a4. Approximate values of actual heat dissipation (hereinafter referred to as “approximate heat dissipation”) ω _so are respectively calculated. In the formula (28) and Equation (29), K _I is the same coefficient and integration coefficient used for integral control or the like. AFMmto is a value calculated by the following equation (30) (that is, from the map of FIG. 18) based on the sum of the approximate heat dissipation amounts ω _bo and ω _so calculated by the observer.

Therefore, in the observer, the output value AFMmto of the air flow meter 41 calculated based on the output value AFM of the air flow meter 41 and the approximate heat dissipation ω _bo and ω _so calculated by the observer (hereinafter referred to as “approximate output value”). (AFM-AFMmto) is accumulated over time. The cumulative value of the deviation between the output value AFM and the approximate output value AFMmto changes based on the error between the actual throttle passage air flow rate and the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10. The absolute value of the cumulative value of the deviation increases as the absolute value of increases. Therefore, the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10, adding the value obtained by multiplying the coefficient K _I to the accumulated value of the deviation between the calculated approximate output value AFMmto the output value AFM and observer Thus, an error between the actual throttle passing air flow rate and the estimated value mt of the throttle passing air flow rate calculated by the throttle model M10 can be compensated to some extent. Thus, the observer, the actual throttle-passing air flow rate substantially equal to the throttle-passing air flow rate approximate heat dissipation omega _bo based on, for omega _so is calculated, approximating the heat radiation amount calculated by the observer omega _bo, omega _so the The actual heat radiation amount from the bobbin portion 41a3 and the support portion 41a4 is almost equal.

In order to calculate the approximate heat radiation amounts ω _bo and ω _so from the estimated value mt of the throttle passage air flow rate and the output value AFM of the air flow meter 41 by mounting the observer in the control device of the internal combustion engine, the equation (28 ) And equation (29) need to be discretized. When the equations (28) and (29) are discretized using the time t and the calculation interval Δt, the following equations (31) and (32) are obtained, respectively. In addition, ΔAFMINT (t) in Expression (31) and Expression (32) is a cumulative value of AFM (t) −AFMmto (t) at time t, and is expressed by Expression (33) below.

In the observer implemented in this way, the approximate heat dissipation amounts ω _bo (t) and ω _so (t) from the bobbin portion 41a3 and the support portion 41a4 at time t, the time calculated by the equation (8) of the throttle model M10. The estimated value mt (t) of the air flow rate through the throttle at t, the output value AFM (t) of the air flow meter 41 at time t, and the approximate output value AFMmto (t) calculated by the observer at time t are expressed by Equation (31), Substituting into Equations (32) and (33), approximate heat dissipation amounts ω _bo (t + Δt) and ω _so (t + Δt) at time t + Δt are calculated. Then, based on these approximate heat dissipation amounts ω _bo (t + Δt) and ω _so (t + Δt), an approximate output value AFMmto (t + Δt) at time t + Δt is calculated by equation (30), and the approximate output value AFMmto (t + Δt) at time t + Δt. Is used for the subsequent calculation of the approximate heat dissipation ω _bo and ω _so . Then, by repeating such calculation, the approximate heat dissipation amounts ω _bo and ω _{so at} an arbitrary time t are calculated from the estimated value mt of the air flow rate through the throttle and the output value AFM of the air flow meter 41.

FIG. 21 is a time chart showing an enlarged transition of the parameter relating to the flow rate at times t _{0 to} t ₁ in FIG. As can be seen from FIG. 21, the approximate output value AFMmto (short broken line in FIG. 21) calculated by the observer is almost the same value as the output value AFM of the air flow meter 41. As described above, in calculating the corrected output value AFMmte, when the throttle valve opening θt enters the map area A, that is, at the time t ₀ , the approximate heat release amount ω _bo of the bobbin portion 41a3 and the support portion An approximate heat release amount ω _{so of} 41a4 is used. Therefore, in the calculation of the corrected output value AFMmte, the initial values of the delayed heat radiation amounts ω _be and ω _se , that is, the delayed heat radiation amounts ω _be (t ₀ ) and ω _se (t ₀ ) at the time t ₀ were calculated by the observer. Approximate heat dissipation ω _bo (t ₀ ) and ω _se (t ₀ ) at time t ₀ (ω _be (t ₀ ) = ω _bo (t ₀ ), ω _se (t ₀ ) = ω _so (t ₀ )). Therefore, fix the output value AFMmte at time t ₀ has the same value as the approximate output value AFMmto of the observer at time t _0.

Then, when the calculation of the corrected output value AFMmte, delay heat dissipation omega _BE, the initial value of omega _se, i.e. lag heat radiation amount omega _BE at time t ₀ (t _0), after ω _se (t ₀₎ is calculated, above expression (26) and lag heat radiation amount at time t ₀ after the equation (27) omega _bE (t), omega _se (t) is calculated, and delay the heat radiation amount omega _bE in these time t ₀ after the time t Based on (t) and ω _se (t), the corrected output value AFMmte (t) at time t after time t ₀ is calculated using the above equation (25). The corrected output value AFMmte calculated in this way changes as shown by the dotted line in FIG. That is, the corrected output value AFMmte is the same value as the approximate output value AFMmto calculated by the observer at time t ₀ , and then gradually shifts to approach the expected output value AFMmt calculated by the AFM model M50. .

The corrected output value AFMmte calculated in this way is a value from which the influence of the estimated value mt of the throttle passage air flow rate before the throttle valve opening θt enters the map area A is removed as described above. . For this reason, the ratio between the output value AFM of the air flow meter 41 and the corrected output value AFMmte in the map area A is the ratio of the actual throttle passing air flow rate in the map area A and the throttle passing air flow rate calculated by the throttle model M10. This is shown relatively accurately, that is, the error rate of the flow coefficient μt of the throttle model M10 in the map area A is shown. Therefore, similarly to the above equation (22), the error rate Err is calculated by the following equation (34) based on the output value AFM of the air flow meter 41 and the corrected output value AFMmte.

That is, the error rate Err is a value obtained by integrating the output value AFM of the air flow meter 41 while the throttle valve opening θt is in a specific map region (map region A in the examples shown in FIGS. 20 and 21), that is, A value obtained by integrating the output value AFM from time t ₀ to t ₁ (area S ₃ + S _{4 in} FIG. 21), the throttle valve opening θt is the specific map region (in the example shown in FIGS. 20 and 21). The value obtained by integrating the corrected output value AFMmte over the map area A), that is, the value obtained by dividing the corrected output value AFMmte by the value (area S _{3 in} FIG. 21) integrated from the time t ₀ to t ₁ . It is equal to ((S ₃ + S ₄ ) / S ₃ ). By multiplying the error rate Err calculated in this way by the map value of the flow coefficient μt in the specific map area (the map area A in the examples shown in FIGS. 20 and 21), the map value of the flow coefficient μt is obtained. Corrected accurately.

In the above embodiment, the map value of the flow coefficient μt corresponding to the map area A where the throttle valve opening θt is θt ₀ to θt ₁ has been described. However, the same applies to other map areas as well. The map value of the flow coefficient μt corresponding to the map area is corrected. Thereby, even during the transient operation of the internal combustion engine, the map value of the flow coefficient μt corresponding to each map area can be corrected accurately.

  By the way, as described above, by using the corrected output value AFMmte, the influence of the estimated value mt of the throttle passage air flow before the throttle valve opening θt enters the map area A is removed. However, the output value AFM of the air flow meter 41 is also affected by the actual throttle passage air flow rate before the throttle valve opening θt enters the map area A, and even if the corrected output value AFMmte is used as described above. The influence of the past actual throttle passage air flow rate cannot be removed from the output value AFM of the air flow meter 41. For this reason, as described above, the error rate of the flow coefficient μt of the throttle model M10 calculated based on the output value AFM and the corrected output value AFMmte of the air flow meter 41 is different from the error rate of the actual flow coefficient μt. Slightly different values.

  Therefore, in the modified example of the above embodiment, the influence of the past actual throttle passage air flow rate is removed from the output value AFM of the air flow meter 41. Hereinafter, a method for accurately calculating the error rate of the flow coefficient μt of the throttle model M10 by removing the influence of the past actual throttle passage air flow rate from the output value AFM of the air flow meter 41 will be described.

  Incidentally, in general, the response of a sensor such as the air flow meter 41 can be separated into an initial value response and a zero response. Here, the initial value response is a transition of the output value of the sensor after time t when it is assumed that the state quantity after a certain time t is zero, and the zero response is a state until a certain time t. It is a transition of the output value of the sensor after time t when it is assumed that the quantity is zero.

FIG. 22 is a diagram showing an initial value response curve and a zero response curve for the output value AFM of the air flow meter 41. Here, Φ ₁ indicates an initial value response curve (broken line) with respect to time 0, and the output value of the air flow meter 41 when it is assumed that the actual air flow rate through the throttle is zero after time 0. It shows the transition. Also, Φ ₂ indicates a zero response curve (one-dot chain line) with respect to time 0. When it is assumed that the actual flow rate of air passing through the throttle up to time 0 is zero, that is, from time 0, the air flow meter 41 The transition of the output value of the air flow meter 41 when the measurement is started is shown.

That is, the initial value response curve changes according to the output value of the air flow meter 41 at time 0, that is, according to the actual throttle passage air flow before time 0, while the zero response curve changes from time 0. It changes according to the actual throttle passage air flow rate at. The output value AFM (t) of the air flow meter 41 at each time t after time 0 is a value Φ ₁ (t) at the time t of the initial value response curve and a value Φ ₂ (t) at the time t of the zero response curve. (AFM (t) = Φ ₁ (t) + Φ ₂ (t)). The relationship between the output value AFM of the air flow meter 41, the value Φ _{1 at} each time of the initial value response curve, and the value Φ _{2 at} each time of the zero response curve is a reference time (see FIG. 22). In the example shown, the time 0) is established at any time.

  As described above, the zero response curve changes in accordance with the actual throttle passage air flow rate after a predetermined time. Therefore, from the zero response curve of the output value AFM of the air flow meter 41 based on the time of entry into the specific map area, The effect of the actual throttle passage air flow before entering that particular map area has been eliminated. Therefore, by obtaining such a zero response curve and comparing it with the zero response curve for the corrected output value AFMmte (more precisely, the curve corresponding to the zero response curve for the corrected output value AFMmte), as described later. The error rate of the flow coefficient μ can be accurately calculated by removing the influence of the actual throttle passage air flow before entering the specific map area.

However, since it is difficult to calculate the zero response based on the time t ₀ , in this embodiment, the initial value response curve based on the time t ₀ from the output value AFM of the air flow meter 41 at each time. The zero response is calculated by subtracting the value Φ ₁ (Φ ₂ (t) = AFM (t) −Φ ₁ (t)).

Here, a method of calculating the value Φ _{1 at} each time of the initial value response curve with the time t ₀ as a reference will be described. As described above, the value Φ _{1 at} each time of the initial value response curve indicates the transition of the output value of the air flow meter 41 when it is assumed that the actual throttle passage air flow rate is zero after the time t ₀ . The estimated value mt of the air flow rate through the throttle in the above equations (13) and (14) can be obtained by setting it to zero. That is, the heat dissipation amounts ω _bi and ω _si for the initial value response curve are expressed by the following equations (35) and (36). In this case, the heat radiation amount omega _bi (t ₀₎ at time t _0, omega _si (t ₀₎ is an approximation of actual heat radiation amount at time t ₀ approximate heat radiation amount _{ω bo (t 0), ω} so ( t ₀ ) is used.

Then, based on the heat radiation amounts ω _bi and ω _si calculated by the equations (35) and (36), the value at each time of the initial value response curve from the map similar to the map shown in FIG. 18 by the equation (37). Φ ₁ is calculated. Equations (35) and (36) are discretized for use by ECU 31.

By subtracting the value Φ ₁ at each time of the initial value response curve based on the time t ₀ calculated in this way from the output value AFM of the air flow meter 41, the time t ₀ for the output value of the air flow meter 41 is obtained. A value Φ _{2 at} each time of the zero response curve with reference to is calculated.

Since the corrected output value AFMmte also uses the approximate heat dissipation ω _bo and ω _so at time t ₀ , by subtracting the value Φ ₁ at each time of the initial value response curve described above from the corrected output value AFMmte, A value at each time of the zero response curve for the corrected output value AFMmte is calculated. For this reason, in this embodiment, when calculating the error rate Err of the flow coefficient μt of the throttle model M10, the time t ₀ regarding the output value AFM of the air flow meter 41, the corrected output value AFM mte, and the output value of the air flow meter 41. Is calculated by the following equation (38) using the value Φ _{1 at} each time of the initial value response curve with reference to.

As described above, in calculating the error rate of the flow coefficient μt, by using the corrected output value AFMmt, the influence of the estimated value mt of the past throttle passage air flow rate existing in the expected output value AFMmt calculated by the AFM model M50 is obtained. That is, the component included in the predicted output value AFMmt is removed from the predicted output value AFMmt due to the influence of the flow rate of air passing through the throttle a predetermined time ago. Further, by subtracting the value Φ ₁ at each time of the initial value response curve from the output value AFM and the corrected output value AFMmte of the air flow meter 41, past actual throttle passage air existing in the output value AFM of the air flow meter 41 and the like. The influence of the flow rate is removed, that is, the component included in the output value AFM of the air flow meter can be removed from the output value of the air flow meter due to the influence of the air flow rate through the throttle before a predetermined time.

  For this reason, the error rate Err calculated by the above equation (38) is the actual throttle passing air flow rate when the throttle valve opening θt is within the predetermined map region and the throttle passing air flow rate calculated by the throttle model M10. Since the calculation is based on the estimated value, the error rate of the flow coefficient when the throttle valve opening degree θt is within a predetermined map region can be accurately calculated.

Next, the correction operation of the AFM model M50 will be described. In describing the correction operation of the AFM model M50, first, the time constants τ _B and τ _S of the first-order lag in heat radiation from the bobbin portion 41a3 and the support portion 41a4 used in the model calculation formulas (13) and (14) of the AFM model M50. The necessity of correcting the will be described.

With the operation of the internal combustion engine, deposits adhere to the bobbin portion 41a3 and the support portion 41a4 of the air flow meter 41. When deposits are attached in this manner, the heat capacities of the bobbin portion 41a3 and the support portion 41a4 increase, and a delay occurs in heat dissipation from the bobbin portion 41a3 and the support portion 41a4. For this reason, an error occurs in the time constants τ _B and τ _S used in the equations (13) and (14) of the AFM model M50 with the deposits attached to the bobbin portion 41a3 and the support portion 41a4. Thus, when an error occurs in the time constants τ _B and τ _S , an error occurs in the in-cylinder charged air amount Mc calculated by the air model M2.

  FIG. 23 is a time chart of the intake pipe pressure calculated by the model blocks M2 ′, M2 ″, M2 ′ ″. As described with reference to FIG. 13, in the air model M2, the air flow meter calculated in the model block M2 ′ ″ is added to the intake pipe pressure Pmf (solid line in the figure) after a predetermined time calculated in the model block M2 ′. The difference between the estimated value Pmafm of the intake pipe pressure based on the output of the engine (broken line in the figure) and the estimated value Pmmdl of the intake pipe using the AFM model M50 calculated by the model block M2 ″ (dashed line in the figure) (Α in the figure) is to be added.

Here, for example, when the time constants τ _b and τ _{s of the} AFM model M50 do not correspond to the actual first-order lag in the air flow meter 41 and are too large, intake using the AFM model M50 calculated in the model block M2 ″ The estimated value of the in-pipe pressure changes as shown by a two-dot chain line indicated by Pmmdl ′ in FIG. That is, as described above, the difference α is calculated based on the current intake pipe pressure calculated by the air model and the air flow meter 41 if the time constants τ _b and τ _s of the AFM model M50 are appropriate values. Although the error from the intake pipe pressure is accurately represented, the error is not accurately represented unless the time constants τ _b and τ _{s of the} AFM model M50 are appropriate values. Accordingly, if the cylinder charge air amount Mc is calculated based on the intake pipe pressure Pmf after a predetermined time plus the difference α when the time constants τ _b and τ _{s of} the AFM model M50 are not appropriate, the cylinder charge air The amount Mc cannot be calculated appropriately. For this reason, in order to appropriately calculate the in-cylinder charged air amount Mc in the air model M2, the time constants τ _b and τ _s used in the AFM model M50 are appropriately set so as to correspond to the primary delay in the actual air flow meter 41. Must be set to a value.

Next, a basic concept for correcting the model calculation formula of the AFM model M50, particularly the time constants τ _b and τ _s will be described. As described above, the corrected output value AFMmte used in the correction operation of the throttle model M10 is a value calculated using the AFM model M50, and is naturally influenced by the time constants τ _b and τ _s .

  FIG. 24 is a time chart of values of various parameters regarding the air flow rate. In FIG. 24, in particular, there is an error between the actual throttle passage air flow rate and the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10, and the estimation of the throttle passage air flow rate is larger than the actual throttle passage air flow rate. The transition of the values of various parameters when the value mt is smaller is shown.

24A shows a case where the actual throttle passage air flow rate is gradually increasing (that is, during acceleration of the internal combustion engine), and FIG. 24B shows a case where the actual throttle passage air flow rate is gradually reduced. (Ie, when the internal combustion engine is decelerating). In the figure, AFM is an output value of the air flow meter 41, and AFMmte is a correction calculated when the time constants τ _b and τ _s used in the AFM model M50 are appropriate values corresponding to the actual first-order delay. The output value, AFMmte ′, is a corrected output value calculated when the time constants τ _b and τ _s used in the AFM model M50 are smaller than appropriate values, and AFMmte ″ is used in the AFM model M50. The corrected output values calculated when the time constants τ _b and τ _s are larger than appropriate values are shown.

When the throttle passage air flow rate gradually increases as shown in FIG. 24A, the time constants τ _b and τ _s used in the AFM model M50 are smaller than appropriate values. The corrected output value AFMmte tends to increase. The reason is as follows. That is, the corrected output value AFMmte is basically calculated by the AFM model M50 based on the estimated value mt of the throttle passage air flow rate calculated by the throttle model M10. When the time constants τ _b and τ _s used in the AFM model M50 are smaller than appropriate values, the corrected output value AFMmte is calculated as a response delay smaller than the actual response delay of the air flow meter 41. Will be. When the actual throttle passage air flow rate is increasing, when the corrected output value AFMmte is calculated on the assumption that the response delay of the air flow meter 41 is smaller than the actual response, the increasing throttle passage air flow rate In response to the estimated value mt quickly, the corrected output value AFMmte immediately rises, changes like AFMmte ′ indicated by a one-dot chain line in FIG. 24A, and becomes a large value. In particular, in the example shown in FIG. 24A, when the corrected output value is a large value such as AFMmte ′, the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte ′ is small. Become.

On the other hand, when the throttle passage air flow rate gradually increases as shown in FIG. 24A, the time constants τ _b and τ _s used in the AFM model M50 are larger than appropriate values. Then, the corrected output value AFMmte tends to be small. That is, if the time constants τ _b and τ _s used in the AFM model M50 are large, the corrected output value AFMmte is calculated as a response delay larger than the actual response delay of the air flow meter 41. When the actual throttle passage air flow rate is increasing, if the corrected output value AFMmte is calculated on the assumption that the response delay of the air flow meter 41 is larger than the actual one, the increased throttle passage air flow rate In response to the estimated value mt slowly, the corrected output value AFMmte rises with a large delay, transitions like AFMmte ″ indicated by a two-dot chain line in FIG. 24A, and becomes a small value. In particular, in the example shown in FIG. 24A, when the corrected output value is a small value such as AFMmte ″, the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte ″ is large. It will be a thing.

Further, as shown in FIG. 24B, when the throttle passage air flow rate is gradually decreasing, the time constants τ _b and τ _s used in the AFM model M50 are smaller than appropriate values. The corrected output value AFMmte tends to be small. That is, as described above, when the time constants τ _b and τ _s are small, the corrected output value AFMmte is calculated as a response delay smaller than the actual response delay of the air flow meter 41, and therefore the throttle that is decreasing In response to the estimated value mt of the passing air flow rate quickly, the corrected output value AFMmte immediately falls, transitions like AFMmte ′ indicated by a one-dot chain line in FIG. 24B, and becomes a small value. In particular, in the example shown in FIG. 24B, when the corrected output value becomes a small value such as AFMmte ′, the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte ′ is large. Become.

On the other hand, as shown in FIG. 24B, when the throttle passage air flow rate is gradually decreasing, the time constants τ _b and τ _s used in the AFM model M50 are larger than appropriate values. The corrected output value AFMmte tends to increase. That is, as described above, when the time constants τ _b and τ _s are large, the corrected output value AFMmte is calculated assuming that a response delay larger than the actual response delay of the air flow meter 41 is generated. Slowly responding to the estimated value mt of the passing air flow rate, the corrected output value AFMmte descends with a large delay, transitions like AFMmte ″ indicated by a two-dot chain line in FIG. 24B, and becomes a large value. In particular, in the example shown in FIG. 24B, when the corrected output value is a large value such as AFMmte, the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte ″ is small. Become.

In summary, when the time constants τ _b and τ _s used in the AFM model M50 are smaller than appropriate values, the corrected output value AFMmte becomes a large value when the throttle-passing air flow rate is increasing (FIG. 24). On the other hand, when the throttle passage air flow rate is decreasing, the corrected output value AFMmte becomes a small value (see AFMmte ′ in FIG. 24B). Conversely, when the time constants τ _b and τ _s used in the AFM model M50 are larger than the optimum values, the corrected output value AFMmte becomes a small value when the throttle passage air flow rate is increasing (FIG. 24A). On the other hand, when the throttle passage air flow rate decreases, the corrected output value AFMmte becomes a large value (see AFMmte ″ in FIG. 24B).

In particular, as shown in FIG. 24, when the estimated value mt of the throttle passage air flow calculated by the throttle model M10 is smaller than the actual throttle passage air flow, the time constants τ _b and τ _s are optimum values. Is smaller than the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte when the throttle passage air flow rate is increasing, and conversely when the throttle passage air flow rate is decreasing, the difference ΔAFM. Becomes larger. On the other hand, when the time constants τ _b and τ _s are larger than the optimum values, the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte becomes large when the throttle passage air flow rate is increasing. This difference ΔAFM decreases when the throttle-passing air flow rate increases. Further, when the time constants τ _b and τ _s are optimum values, the air flow meter 41 is used even when the throttle passage air flow rate is increasing or when the throttle passage air flow rate is decreasing. The difference ΔAFM between the output value AFM and the corrected output value AFMmte is almost the same value.

Therefore, in the present invention, when the throttle passage air flow rate is increasing and when the throttle passage air flow rate is increasing, the time constants τ _b and τ _s used in the AFM model M50 are optimized. When the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte is different from time to time, the time constants τ _b and τ _s are corrected so that the difference ΔAFM becomes the same value. .

In particular, in the present embodiment, as described above, the time constants τ _b and τ _s are calculated by the equations (16) and (17), respectively. Of these, the constants k _b and k _s are corrected. . The constants k _b and k _s are determined for each fixed engine operating region and are stored in the ROM 34 of the ECU 31 as a map similar to the map shown in FIG. Specifically, the map of the constants k _b and k _s divides the throttle valve opening θt into a plurality of map areas, and one constant k _b and k _s map for each map area of the throttle valve opening θt. Value is set. Therefore, when the constants k _b and k _s are corrected, the map values of the constants k _b and k _s corresponding to the map area are corrected for each map area of the throttle valve opening θt.

Therefore, in the present embodiment, based on the output value AFM and the corrected output value AFMmte of the air flow meter 41 when the throttle valve opening degree θt is in a certain map area, the constants k _b and k _s corresponding to the map area. I am going to modify the map value.

Specifically, the constants k _b and k _s of the time constants τ _b and τ _s are corrected as follows. First, when the throttle passage air flow rate is increasing, the time when the throttle valve opening θt enters a certain map region is t ₀ , and the time when the throttle valve opening θt leaves is t ₁ , from time t ₀ to time t ₁ . The integrated value of the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte (that is, the solid line of the output value AFM indicated by hatching in FIG. 25, the broken line of the corrected output value AFMmte, the straight line at time t ₀ , and the time ΔAMFave obtained by dividing the area surrounded by the straight line t ₁ by the time from time t ₀ to time t ₁ (hereinafter referred to as “average error”) is calculated by the following equation (39). The

  Similarly, when the throttle-passing air flow rate is decreasing, the output value AFM and the corrected output value AFMmte of the air flow meter 41 from when the throttle valve opening degree θt enters the certain map area until it leaves. An average error ΔAFMave obtained by dividing the integral value of the difference ΔAFM with the time from when the throttle valve opening θt enters the certain map area until it leaves is calculated by the above equation (39).

In this way, the average error ΔAFMave corresponding to a certain map area is calculated at least once each time the throttle passage air flow rate increases and decreases. When the average error ΔAFMave is positive, that is, when the output value of the air flow meter 41 is larger than the corrected output value AFMmte, the average error calculated when the throttle passage air flow rate is increasing (hereinafter referred to as “when increasing”). Time constant τ _b and τ _s when the average error ΔFMavei is smaller than the average error calculated when the throttle passage air flow rate is decreasing (hereinafter referred to as “average error ΔAFMave at the time of decrease”). The values of the constants k _b and k _s are corrected so as to increase, and the constants k _b and τ _s are reduced so that the time constants τ _b and τ _s decrease when the average error ΔAFMavei at the time of increase is larger than the average error ΔAFMaveed at the time of decrease Correct the value of k _s .

On the other hand, when the average error ΔAFMave is negative, that is, when the output value of the air flow meter 41 is smaller than the corrected output value AFMmte, the time constant τ _b and the average error ΔAFMavei at the time of increase are larger than the average error ΔAFMave at the time of decrease. The values of the constants k _b and k _s are corrected so that τ _s becomes large, and the constant k so that the time constants τ _b and τ _s become small when the average error ΔAFMavei when increasing is smaller than the average error ΔAFMaveed when decreasing Correct the values of _b and k _s .

Furthermore, when the average error ΔAFMavei at the time of increase and the average error ΔAFMaved at the time of decrease are substantially the same, the time constants τ _b and τ _s of the AFM model M50 are considered to be appropriate values. The time constants τ _b and τ _s are not corrected.

The difference Δid between the average error ΔAFMavei at the time of increase and the average error ΔAFMaved at the time of decrease represents the magnitude of deviation of the time constants τ _b and τ _s used in the AFM model M50, and this difference Δid is large. The deviation of the time constants τ _b and τ _s from appropriate values is large.

Therefore, in the present embodiment, when correcting the time constants τ _b and τ _s , that is, when correcting the values of the constants k _b and k _s , the average error ΔAFMavei at the time of increase and the average error ΔAFMaved at the time of decrease. The correction rate Mdf of the constants k _b and k _s is determined according to the difference Δid between and the constants k _b and k _{s so that} the correction rate of the constants k _b and k _s increases as the difference Δid increases. Note that the relationship between the difference Δid of the average error ΔAFMave and the correction amount of the time constants τ _b and τ _s can be calculated by simulation or can be calculated in advance by experiment, and based on this, the average error ΔAFMave can be calculated. The relationship between the difference Δid and the correction rate Mdf is stored in the ROM 34 of the ECU 31 as a map. The correction rate Mdf is calculated from this map based on the difference Δid of the average error ΔAFMave calculated as described above during engine operation, and the current constants k _b and k _s are multiplied by the correction rate Mdf. Are the new constants k _b and k _s , respectively (k _b = Mdf · k _b , k _s = Mdf · k _s ).

In this way, the constants k _b and k _s of the time constants τ _b and τ _s of the AFM model M50 are corrected so that the time constants τ _b and τ _s correspond to the first order delay in the actual air flow meter 41. It can be corrected to an appropriate value.

  Note that whether or not the throttle passage air flow rate is increasing or decreasing is, for example, an air flow meter between the time when the throttle valve opening θt enters a certain map region and then leaves. This is determined by an increase or decrease in the output value AFM of 41 or the air flow rate mt passing through the throttle. In addition, when the time from when the throttle valve opening θt enters a certain map area to when it is longer than a predetermined time, the degree of increase or decrease of the throttle passage air flow rate per unit time is small. The error ΔAFMave may not be calculated.

  FIG. 26 is a flowchart showing a control routine for calculating / saving control of the average error ΔAFMavei at the time of increase or the average error ΔAFMaved at the time of decrease. The control routine shown in FIG. 26 is performed by interruption at regular time intervals.

  First, in step 101, it is determined whether or not the throttle valve opening θt is in a predetermined map region. When the throttle valve opening θt has entered the predetermined map area, it is determined in step 101 that the throttle valve opening θt is in the predetermined map area, and the routine proceeds to step 102 and step 103. In step 102 and step 103, the output value AFM of the air flow meter 41 is detected and the corrected output value AFMmte is calculated. Next, in step 104, the intrusion flag X is set to 1. The intrusion flag X is a flag that is set to 1 when the throttle valve opening degree θt is in a predetermined map region, and is set to 0 in other cases. The steps 101 to 104 are repeated while the throttle valve opening θt is within the predetermined map region.

  Thereafter, when the throttle valve opening θt departs from the predetermined map area, the routine proceeds from step 101 to step 105. In step 105, it is determined whether or not the intrusion flag X is 1. Immediately after the throttle valve opening θt departs from the predetermined map area, the intrusion flag X remains 1, and in this case, the routine proceeds to step 106. In step 106, the average error ΔAFMave is calculated by the equation (39) based on the output value AFM and the corrected output value AFMmte detected and calculated in steps 102 and 103.

  Next, at step 107, it is determined whether or not the average error ΔAFMave calculated at step 106 is greater than or equal to a predetermined value P. The predetermined value P is a relatively small value, and is such a value that the influence of the noise of the air flow meter 41 becomes large if the average error ΔAFMave is smaller than that. If it is determined in step 107 that the average error ΔAFMave is smaller than the predetermined value P, the control routine is terminated. On the other hand, if it is determined that the average error ΔAFMave is equal to or greater than the predetermined value P, the routine proceeds to step 108. In step 108, it is determined whether the output value AFM of the air flow meter 41 has increased or decreased while the throttle valve opening θt is within the predetermined map region.

  Next, in step 109, if it is determined in step 108 that the output value AFM of the air flow meter 41 has increased while the throttle valve opening θt is within the predetermined map area, the average calculated in step 106 is determined. The error ΔAFMave is stored in the ROM 34 of the ECU 31 as the average error ΔAFMavei at the time of increase. On the other hand, in step 108, it is determined that the output value AFM of the air flow meter 41 has decreased while the throttle valve opening θt is within the predetermined map area. In the case where it is determined, the average error ΔAFMave calculated in step 106 is stored in the ROM 34 of the ECU 31 as the average error ΔAFMavei at the time of decrease. Next, at step 110, the intrusion flag is set to 0 and the control routine is terminated. In the next routine, it is determined in step 105 that the intrusion flag is not 1, and the control routine is terminated.

27 corrects the constants k _b and k _s of the time constants τ _b and τ _s based on the average error ΔAFMavei at the time of increase and the average error ΔAFMave at the time of decrease calculated and stored by the control routine shown in FIG. It is a flowchart which shows a control routine. The control routine shown in FIG. 27 is started after ΔAFMavei and ΔAFMaveed are each calculated by the control routine shown in FIG. 26 one by one after the previous execution of this control routine.

First, at step 121, the average error ΔAFMavei at the time of increase and the average error ΔAFMaved at the time of decrease calculated in step 109 of ΔAFMave calculation / storage control shown in FIG. 26 are acquired. Next, at step 122, the difference Δid is obtained by subtracting the decrease average error ΔAFMaveed acquired at step 121 from the average error ΔAFMavei at the time of increase acquired at step 121 (or the average value when there are a plurality of times). Calculated (Δid = ΔAFMavei−ΔAFMaveed). Next, at step 123, a correction rate Mdf is calculated by a map based on Δid calculated at step 122. Next, at step 124, the current constants k _b and k _s of the time constants τ _b and τ _s of the AFM model M50 are multiplied by the correction rate Mdf calculated at step 123 to obtain new constants k _b and k _s . (K _b = Mdf · k _b , k _s = Mdf · k _s ).

  In the above embodiment, the average error ΔAFMave is a value based on the difference between the output value AFM of the air flow meter 41 and the corrected output value AFMmte, but the difference between the output value AFM of the air flow meter 41 and the expected output value AFMmt is used. It is good also as a value based on.

In the above embodiment, the time is based on the difference ΔAFM between the output value AFM of the air flow meter 41 and the corrected output value AFMmte when the throttle passage air flow rate is increasing and when the throttle passage air flow rate is decreasing. Although the constants τ _b and τ _s are to be corrected, the present invention is not limited to this. For example, the difference between the output value AFM of the air flow meter 41 and the corrected output value AFMmte when the internal combustion engine is accelerated and when the internal combustion engine is decelerated, etc. The time constant may be corrected based on the difference between these output values when operating parameters other than the throttle passing air flow rate are changed so that the throttle passing air flow rate changes.

It is a figure which shows the whole internal combustion engine carrying the control apparatus of this invention. It is a schematic perspective view of the air flow meter shown in FIG. It is an expansion perspective view of the heat ray measuring part of the air flow meter shown in FIG. It is a figure which shows the relationship between the output voltage of an airflow meter, and an airflow passage air flow rate. It is a block diagram of an air model. It is a figure which shows the relationship between a throttle-valve opening degree and a flow coefficient. It is a figure which shows the relationship between a throttle-valve opening degree and opening cross-sectional area. It is a figure which shows function (PHI) (Pm / Pa). It is a figure which shows the basic concept of a throttle model. It is a figure which shows the basic concept of an intake pipe model. It is a figure which shows the basic concept of an intake valve model. It is a figure regarding the definition of the cylinder filling gas amount and the cylinder intake gas amount. FIG. 6 is a block diagram of an air model different from the air model shown in FIG. 5. It is a figure which shows the relationship between accelerator pedal depression amount and target throttle valve opening. It is a figure which shows the relationship between a throttle passage air flow rate and complete heat dissipation. It is a figure which shows the relationship between the sum of delayed heat dissipation, and the estimated output value of an airflow meter. It is a figure which shows correction operation of a throttle model. It is a time chart with the value of various parameters about a throttle valve opening and an air flow rate. It is a map of a flow coefficient determined based on the throttle valve opening. It is a time chart with the value of various parameters about a throttle valve opening and an air flow rate. It is the figure which expanded a part of FIG. It is a figure for demonstrating an initial value response and a zero response. It is a time chart of the pressure in an intake pipe. It is a time chart of the output value and corrected output value of an air flow meter. It is a time chart with the value of the various parameters regarding an air flow rate. It is a flowchart which shows the control routine of calculation / preservation | save control of the average error at the time of increase, or the average error at the time of decrease. It is a flowchart showing a control routine for correcting the constants k _b and k _s of the time constant tau _b and tau _s.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Engine main body 5 Combustion chamber 6 Intake valve 7 Intake port 8 Exhaust valve 11 Fuel injection valve 14 Intake branch pipe 15 Surge tank 16 Intake pipe 19 Throttle valve 23 Intake pipe part

Claims (5)

An air flow meter provided in the engine intake passage;
A throttle passage air flow rate estimating means for estimating a throttle passage air flow rate;
An air flow meter model that calculates an output value that would be output by the air flow meter using an air flow meter model calculation formula based on the estimated value of the throttle flow air flow rate estimated by the throttle passage air flow rate estimation means. And
In the control device for an internal combustion engine that controls the internal combustion engine using an actual measurement value of the air flow meter and an expected output value of the air flow meter calculated by the air flow meter model,
Error calculating means for calculating an error between an expected output value of the air flow meter calculated by the air flow meter model and an actual measurement value of the air flow meter;
The air flow meter model calculation based on the error calculated by the error calculation means when the throttle passage air flow rate is increasing and the error calculated by the error calculation means when the throttle passage air flow rate is decreasing. A control device for an internal combustion engine, further comprising model calculation formula correcting means for correcting the formula.   The model calculation formula correction means corrects the air flow meter model calculation formula so that the error calculated by the error calculation means is equal between when the throttle passage air flow rate is increasing and when it is decreasing. The control apparatus for an internal combustion engine according to claim 1. An air flow meter provided in the engine intake passage;
A throttle passage air flow rate estimating means for estimating a throttle passage air flow rate;
An air flow meter model that calculates an output value that would be output by the air flow meter using an air flow meter model calculation formula based on the estimated value of the throttle flow air flow rate estimated by the throttle passage air flow rate estimation means. And
In the control device for an internal combustion engine that controls the internal combustion engine using an actual measurement value of the air flow meter and an expected output value of the air flow meter calculated by the air flow meter model,
Error calculating means for calculating an error between an expected output value of the air flow meter calculated by the air flow meter model and an actual measurement value of the air flow meter;
The air flow meter model calculation formula is corrected based on the error calculated by the error calculation means when the internal combustion engine is accelerating and the error calculated by the error calculation means when the internal combustion engine is decelerating. A control device for an internal combustion engine, further comprising model calculation formula correcting means.   The model calculation formula correcting means corrects the air flow meter model calculation formula so that the error calculated by the error calculating means is equal when the internal combustion engine is accelerating and when decelerating. The control device for an internal combustion engine according to claim 3.   5. A time constant relating to a response time of the air flow meter with respect to an actual change in the air flow rate through the throttle is used in the air flow meter model calculation formula, and the model calculation formula correcting means corrects the time constant. The control device for an internal combustion engine according to any one of the above.

Images