JPH05180044A - Air-fuel ratio control method for internal combustion engine - Google Patents

Abstract

PURPOSE:To promptly attain convergence to a target value by modeling a detecting response delay of an air-fuel ratio sensor, estimating air-fuel ratio of each cylinder in good accuracy, and feedback-controlling the air-fuel ratio to the target value based on an estimated value. CONSTITUTION:A single air-fuel ratio sensor 40 is provided in an exhaust system collective part 24 of a multicylinder internal combustion engine 10, and air-fuel ratio of a mixture input from an output of this air-fuel ratio sensor 40 is detected and convergency-controlled to a target value. Here in a control unit 42, an output value of the sensor 40 is considered to be formed of a weighted mean value, multiplying a combustion record of each cylinder by predetermined weight, to construct a model of describing action of an exhaust system. A condition equation and an output equation with air-fuel ratio of each cylinder serving as an internal condition variable are obtained, and an observer for observing an internal condition is constructed to obtain an output of this observer. Next, air-fuel ratio of each cylinder is estimated from, the observer output, and this time correction value is obtained by multiplying ratio of an estimated value of the air-fuel ratio to its target value by a correction value before a specific period, to immediately convergency-control the air-fuel ratio of each cylinder to the target value.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling an air-fuel ratio of an internal combustion engine, and more specifically, to an air-fuel ratio of an internal combustion engine for quickly controlling the air-fuel ratio of each cylinder of a multi-cylinder internal combustion engine to a target value. Control method

[0002]

2. Description of the Related Art An exhaust gas system of a multi-cylinder internal combustion engine is provided with an air-fuel ratio sensor consisting of an oxygen concentration detecting element to detect the air-fuel ratio of the input fuel, and the fuel supply amount is feedback-controlled to a target value according to the detected value. This is often done, and an example thereof is the technique described in JP-A-59-101562.

[0003]

In the air-fuel ratio control of such a multi-cylinder internal combustion engine, it is common practice to perform PID control of the air-fuel ratio of each cylinder in accordance with the deviation from the target value. Sex is not always satisfactory.
That is, if an air-fuel ratio sensor is provided for each cylinder, the air-fuel ratio of each cylinder can be accurately detected. However, due to cost and durability problems, it is normal for each cylinder to be output from the output of a sensor installed in the exhaust system. The air-fuel ratio of is estimated. Therefore, the air-fuel ratio of each cylinder cannot be accurately known, and the gain cannot be increased too much to prevent hunting.
The control convergence was not very good.

Therefore, an object of the present invention is to solve the above-mentioned drawbacks and to accurately separate and extract the air-fuel ratio of each cylinder from the output of a single air-fuel ratio sensor arranged in the exhaust system collecting portion of a multi-cylinder internal combustion engine. In addition, it is possible to provide an air-fuel ratio control method for an internal combustion engine that performs dead beat control so that the air-fuel ratio of each cylinder immediately converges to a target value.

Furthermore, more generally, even in the case where a sensor is provided for each cylinder to detect the air-fuel ratio, dead-beat control is performed so that the air-fuel ratio of each cylinder immediately converges to the target value in the next control cycle. An object is to provide an air-fuel ratio control method for an internal combustion engine.

[0006]

In order to solve the above-mentioned problems, an air-fuel ratio control method for an internal combustion engine according to the present invention is arranged in an exhaust system collecting portion of a multi-cylinder internal combustion engine, as set forth in, for example, claim 1. In the one that controls the air-fuel ratio to the target value by detecting the air-fuel ratio of the air-fuel mixture input from the output of the single air-fuel ratio sensor, the output value of the sensor is given a predetermined weight to the combustion history of each cylinder. A model that describes the behavior of the exhaust system, which is regarded as consisting of the weighted average value multiplied, is constructed, the state equation and the output equation in which the air-fuel ratio of each cylinder is an internal state variable is obtained, and an observer for observing the internal state is obtained. Is calculated, the output is obtained, the air-fuel ratio of each cylinder is estimated from the observer output, the ratio between the estimated value and the target value is calculated, and this ratio is multiplied by the correction value before the specific control cycle. The air-fuel ratio of each cylinder as a correction value for Be dead beat controlled so as to be immediately converged to a target value, it was as constructed.

[0007]

With the above configuration, the air-fuel ratio of each cylinder can be immediately converged to the target value in the next control cycle.

[0008]

Embodiments of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a schematic diagram showing an overall air-fuel ratio detection / control apparatus for an internal combustion engine for realizing the method according to the present invention. In the figure, reference numeral 10 indicates a four-cylinder internal combustion engine, and intake air introduced from an air cleaner 14 arranged at the tip of an intake passage 12 is passed through an intake manifold 18 while its flow rate is adjusted by a throttle valve 16. It is introduced into the first to fourth cylinders. An injector 20 is provided near the intake valve (not shown) of each cylinder to inject fuel. The air-fuel mixture injected and integrated with the intake air is ignited by a spark plug (not shown) in each cylinder and burned to drive a piston (not shown). The exhaust gas after combustion is discharged to the exhaust manifold 22 via an exhaust valve (not shown), is discharged to the outside of the engine while being purified by the three-way catalytic converter 26 via the exhaust pipe 24. Further, the intake passage 12 is provided with a bypass passage 28 which bypasses the throttle valve 16 near the position where the throttle valve 16 is arranged.

A crank angle sensor 34 for detecting a crank angle position of a piston (not shown) is provided in a distributor (not shown) of the internal combustion engine 10, and a throttle opening for detecting an opening of the throttle valve 16 is provided. A degree sensor 36 and an absolute pressure sensor 38 for detecting the intake pressure downstream of the throttle valve 16 by absolute pressure are also provided. Furthermore,
In the exhaust system, an air-fuel ratio sensor 40 including an oxygen concentration detecting element is provided on the downstream side of the exhaust manifold 22 and the upstream side of the three-way catalytic converter 26 to detect the air-fuel ratio of the exhaust gas. The outputs of these sensors 34 and the like are sent to the control unit 42.

FIG. 2 is a block diagram showing details of the control unit 42. The output of the air-fuel ratio sensor 40 is the detection circuit 46.
To the air-fuel ratio (A / F), which has linear characteristics proportional to the oxygen concentration in the exhaust gas over a wide range from lean to rich.
Is detected. The details are described in the application previously proposed by the present applicant (Japanese Patent Application No. 3-169456).
Further description will be omitted. In the following description, this sensor will be referred to as "LAF sensor" (Linear AFB
Sensor). The output of the detection circuit 46 is taken into a microcomputer including a CPU 50, a ROM 52, and a RAM 54 via an A / D conversion circuit 48, and a RAM is stored.
54. Similarly, the analog output of the throttle opening sensor 36 and the like is passed through the level conversion circuit 56, the multiplexer 58 and the second A / D conversion circuit 60, and the digital output of the crank angle sensor 34 is a waveform shaping circuit 62.
After the waveform is shaped by, the output value is counted by the counter 64, and the count value is input into the microcomputer. In the microcomputer, the CPU 50 is R
The feedback control value of the air-fuel ratio is calculated from the detected value according to the command stored in the OM 52, and the injector 20 of each cylinder is driven via the drive circuit 66 and the solenoid valve 70 is driven via the second drive circuit 68. Then, the amount of secondary air passing through the bypass 28 shown in FIG. 1 is controlled.

Next, the operation of this control device will be described.
After all, the operation is to detect the air-fuel ratio, find the deviation from the target value, and control the fuel supply amount (injection amount) so as to eliminate the deviation. It is to accurately estimate the air-fuel ratio of each cylinder from the output of the sensor and to feedback control the air-fuel ratio to the target value based on the estimated detection value as well, so the following description focuses on that point. To do.

In order to accurately separate and extract the air-fuel ratio of each cylinder from the output thereof by disposing only one air-fuel ratio sensor in the exhaust system collecting portion of the multi-cylinder internal combustion engine, first, the detection response of the air-fuel ratio sensor It is necessary to clarify the delay accurately. FIG. 3 shows, as an example of a single-cylinder internal combustion engine, actual measurement data of the response of the air-fuel ratio sensor when the intake air amount is constant and the supplied fuel amount is changed stepwise (“actual measurement value” in the figure). As shown in the figure, when the air-fuel ratio is changed stepwise, the measured value of the LAF sensor output has a delay with respect to the input value, but this delay is due to the chemical reaction of the sensor, so it is not possible to analyze it accurately. Have difficulty. Therefore, the present inventors made a model of this delay as a first-order delay system for the time being, and created a model as shown in FIG. Here, if LAF: LAF sensor output and A / F: input A / F, the state equation can be expressed by the following equation 1.

[0014]

[Equation 1]

When this is discretized with a period ΔT, it becomes as shown in Equation 2. FIG. 5 is a block diagram showing Equation 2.

[0016]

[Equation 2]

Therefore, by using the equation 2, the true air-fuel ratio can be obtained from the sensor output. That is, number 2
By transforming, the value at time k can be calculated back from the value at time k, as shown in expression 4.

[0018]

[Equation 3]

[0019]

[Equation 4]

Specifically, if the expression 2 is expressed by a transfer function using the Z-transform, it becomes as shown in the expression 5. Therefore, the previous air-fuel ratio is estimated in real time by multiplying the inverse transfer function by the sensor output LAF of this time. can do. FIG. 6 shows a block diagram of the real-time A / F estimator. still,
As described above, the response delay of the LAF sensor is caused by a chemical reaction and is difficult to analyze accurately, but it was confirmed that there is a correlation with the engine speed. Therefore, the coefficient of the transfer function is set to be different for each predetermined engine speed that is appropriately set. Therefore, the accuracy of the estimated A / F value can be further improved by switching the A / F estimator, that is, the coefficient of the inverse transfer function for each predetermined engine speed.

[0021]

[Equation 5]

The results of the above simulation are shown in FIG. 3 (“simulation” in the figure) and FIG. 7. As described above, the “actual measurement value” in FIG. 3 is the actual measurement value of the sensor output when a step-like air-fuel ratio input is given.
Here, it can be seen that the measured values and the simulation results (the output obtained by inputting the stepped air-fuel ratio into the model in FIG. 5) substantially correspond to each other. From the above, it can be said that the correctness of artificially modeling the sensor response delay as the first-order delay was verified. FIG. 7 shows a case where the measured value of the sensor output is multiplied by the inverse transfer function to estimate the true air-fuel ratio. In the figure, for example, the true air-fuel ratio at time Ta is 1
It can be estimated to be 13.2 instead of 2.5. It should be noted that the reason that the estimated value of the true air-fuel ratio slightly fluctuates is that there is a slight variation in the actual measurement value of the sensor output.

Next, the case where the air-fuel ratio of each cylinder is separated and extracted based on the true air-fuel ratio obtained as described above will be described.

As described above, when one air-fuel ratio sensor is arranged in the multi-cylinder internal combustion engine, the output thereof shows a value obtained by mixing the detection values of the respective cylinders at the collecting portion of the exhaust system,
It is difficult to find the true air-fuel ratio for each cylinder. Therefore, the A / F of each cylinder cannot be individually controlled to the target value, and some cylinders may be lean or other cylinders may be rich, which causes emission deterioration. In order to solve this, a sensor may be arranged for each cylinder, but this causes an increase in cost and has a problem of durability. Therefore, the inventors of the present invention have been able to model the sensor response delay as a first-order delay. Therefore, one air-fuel ratio sensor arranged in the collecting part of the exhaust system is used in the multi-cylinder, in the case of the embodiment, four cylinders, according to the following method. The air-fuel ratio of the internal combustion engine can be accurately detected for each cylinder. This will be described below.

First, an exhaust system of an internal combustion engine was modeled as shown in FIG. 8 (hereinafter, this model is referred to as an "exhaust manifold model").
Called)). In this exhaust manifold model, the sample time of the discrete system is set to the TDC cycle (the engine speed is 1500 rpm).
Then 0.02 sec). Also, in this exhaust manifold model, since F (fuel) is the controlled variable, the air-fuel ratio is F
/ A.

Here, the inventors considered that the air-fuel ratio (A / F) of the collecting portion of the exhaust system is a weighted average value considering the temporal contribution of the air-fuel ratio of each cylinder. If so, the air-fuel ratio of the collecting portion at the time k can be expressed as in Equation 6.

[0027]

[Equation 6]

That is, the air-fuel ratio of the collecting portion is weighted by C in the past combustion history for each cylinder (for example, the most recently burned cylinder is 40).
%, Before that 30%. ． ． Etc.) can be expressed as the sum of those multiplied. Here, the mixed state of the exhaust gas of each cylinder in the collecting portion varies depending on the operating state of the engine. That is, for example, since the TDC cycle is long in the low engine speed range,
The degree of mixing of exhaust gas from each cylinder is low compared to the high engine speed range. Also, when the load is high, the back pressure is basically large,
Since the exhaust pressure of the exhaust gas increases, the degree of mixing of the exhaust gas from each cylinder is lower than when the load is low. In this way, when the degree of mixing of the exhaust gas of each cylinder is low, it is necessary to give a large weight to the cylinder that most recently burned. Therefore, the weight C is changed depending on the operating state of the engine. Specifically, the weight C is appropriately set using the engine speed and the load as parameters, prepared in the map, and searched for. In the above, #n indicates the cylinder number, and the combustion (ignition) order of the cylinders is 1, 3, 4, 2
And Further, the air-fuel ratio [F / A] here means a true value obtained by correcting the response delay previously obtained by the equation (5).

Based on the above, the equation of state of the exhaust manifold model is as shown in equation 7.

[0030]

[Equation 7]

If the air-fuel ratio of the collecting portion is y (k),
The output equation can be expressed as Equation 8.

[0032]

[Equation 8]

In the above, since u (k) cannot be observed, even if the observer is designed from this equation of state, x
(K) cannot be observed. Therefore, assuming that the air-fuel ratio F / A before 4TDC (that is, the same cylinder) is in a steady operation state where it does not change rapidly, x (k + 1) = x (k-
If it is 3), it becomes like Formula 9.

[0034]

[Equation 9]

Here, simulation results of the exhaust manifold model obtained as described above will be shown. FIG. 9 shows a case where the air-fuel ratio of three cylinders is set to 14.7 and fuel is supplied to only one cylinder at 12.0 in a four-cylinder internal combustion engine. Figure 10
Shows the air-fuel ratio (A / F) of the collecting portion (that is, the position where the air-fuel ratio sensor 40 is arranged in the exhaust manifold pipe 24 in FIG. 1) obtained by the above exhaust manifold model. In FIG. 10, a step-like output is obtained, but if the response delay of the LAF sensor is further taken into consideration here, the sensor output has a waveform shown as “simulation” in FIG. 11. In the figure, the “actual measurement value” is the actual measurement value of the LAF sensor output in the same case.
It can be said that it has been verified that the above exhaust manifold model well models the exhaust system of a multi-cylinder internal combustion engine.

Therefore, the problem of a stationary Kalman filter for observing x (k) in the state equation and the output equation shown in the equation 10 is reduced. When the Riccati equation is solved by using the weighting matrices Q and R as shown in equation 11, the gain matrix K becomes as shown in equation 12.

[0037]

[Equation 10]

[0038]

[Equation 11]

[0039]

[Equation 12]

When A-KC is obtained from this, the result is as shown in the equation 13.

[0041]

[Equation 13]

The general structure of the observer is as shown in FIG. 12, but since there is no input u (k) in this model, the structure is such that only y (k) is input as shown in FIG. When this is expressed by a mathematical expression, it becomes as shown in Expression 14.

[0043]

[Equation 14]

Here, an observer whose input is y (k),
That is, the system matrix of the Kalman filter is expressed as in Expression 15.

[0045]

[Equation 15]

In the model this time, when the element of the weight distribution R of the Riccati equation: the element of Q = 1: 1, the system matrix S of the Kalman filter is given by equation 16.

[0047]

[Equation 16]

Subsequently, a waveform of the cylinder-by-cylinder air-fuel ratio is accurately created on the simulation, and the waveform is input to the exhaust manifold model to obtain the collective portion air-fuel ratio. Input it to the observer and verify that the cylinder-by-cylinder air-fuel ratio is estimated. Moreover, the tendency of the load matrix and the estimated value is considered.

In the model this time, since it is as shown in Expression 17, the weight matrix Q is a diagonal matrix in which all the elements are the same.

[0050]

[Equation 17]

Therefore, what is to be considered is the ratio of the elements of Q and R. Table 1 shows the gains obtained by changing the ratio of the Q and R elements. In addition, the simulation model that combines the observer and exhaust manifold model constructed by using it is shown in Fig. 1.
4 shows. Furthermore, using this model, the cylinder air-fuel ratio is set to 1
FIG. 15 shows the values calculated as the ideal inputs of 2.0, 14.7, 14.7, and 14.7, and Table 2 shows the estimation error of the observer at that time. Furthermore, set the air-fuel ratio to 12.0 ±
0.2, 14.7 ± 0.2, 14.7 ± 0.2, 14.
FIG. 16 shows the values independently calculated (virtual noise) as 7 ± 0.2, and Table 3 shows the estimation error of the observer at that time. 15 (a) to 16 (e), (a) each cylinder A / F (exhaust manifold model input), (b) collecting unit A / F (exhaust manifold model output), (c) Q. Element: R is an observer output when the element is 1:10 (input is shown in (b)), (d) Q element is an observer output when the R element is 1: 1 (input is (b)) , (E) Q element: R element = 10:
It is an observer output when 1 is set (input is shown in (b)).

[0052]

[Table 1]

[0053]

[Table 2]

[0054]

[Table 3]

As shown in FIG. 15, when the air-fuel ratio of each cylinder is constant, it is understood that the larger the weight of Q, the faster the convergence. However, even if Q / R was set to 10 or more, the convergence did not change. FIG. 16 shows the estimated deviation (air-fuel ratio of each cylinder-estimated air-fuel ratio) in time series in FIG. 17, and becomes as shown in FIG. 17, and after the observer converges, the element of Q: the element of R = 10: 1 and 1: 1.
Since there is not much difference between and, considering the resistance to disturbance,
It can be said that the element of Q: the element of R = 1: 1 is better. As described above, the observer using the Kalman filter theory with respect to the input of the air-fuel ratio of the collecting portion enables the air-fuel ratio of each cylinder in the collecting portion to be accurately estimated. The best weight matrix was Q / R = 1-10, but it seems necessary to determine it from the response situation using actual data.

Subsequently, the actual assembly air-fuel ratio data obtained by inputting the actually measured data to the inverse transfer function of the LAF sensor shown above is input to the observer, and the result of estimating the cylinder-by-cylinder air-fuel ratio is shown in FIG. Show. In the figure, (a) LAF sensor output, (b) LAF sensor inverse transfer function output (input is shown in (a)), (c) element of Q: element of R = 1: 10
Observer output at the time of (input is shown in (b)),
(D) Element of Q: element of R = 1: 1 observer output (input is shown in (b)), (e) element of Q: element of R = 10: 1 observer output (input Is shown in (b)). Here, the measurement conditions of the LAF sensor output are
Engine speed = 1500 rpm, intake pressure = −281.9
mmHg, A / F = 12.0 (# 2), 14.7 (#
1, # 3, # 4). Further, since the true value of the actual input A / F is not known, the approximate value in the simulation is [12.0 / 14.7 / 14.7 / 14.7].
Was used. As is clear from the figure, the observer output is 4
It changes in the TDC cycle, and the input A / F is almost estimated. Further, it was confirmed that the Kalman filter can be used to converge in 2 to 8 cycles depending on the setting of the weight matrix.

Next, the case where the air-fuel ratio is controlled to the target value by using the cylinder-by-cylinder air-fuel ratio estimated as described above will be described. FIG. 19 is a block diagram showing an example of control using the PID method. Although it is different from the normal PID control in that it is fed back via the multiplication term, this control method itself is a known one, and as shown in the figure, it is against the target value of the actual air-fuel ratio generated by the input Ti (injection time). The deviation (1-1 / λ) may be calculated for each cylinder, and the target value may be feedback-controlled by multiplying the deviation by the gain KLAF. However, although it depends on a known method, the air-fuel ratio of each cylinder can be accurately detected as described above, and as a result, the air-fuel ratio of each cylinder can be accurately controlled to the target value.

However, in the PID control described above, the convergence cannot be said to be good because the gain cannot be made too large in order to prevent hunting. 20 to 24 show simulation results showing the responsiveness of the PID control of FIG. Figure 20 shows input A
FIG. 21 is a characteristic diagram of the A / F output characteristic when / F is constant (21.0), and FIG. FIG. 22 is another characteristic diagram of the input A / F, FIG. 23 is an output characteristic diagram of the A / F at that time, and FIG. 24 is a characteristic diagram of the gain KLAF corresponding thereto. As is clear from FIG. 23, the convergence cannot be said to be fast. Therefore, the dead beat control that immediately converges to the target value will be described.

As a basic policy, the ratio W of the air-fuel ratio x hat (k) estimated by the observer and the target air-fuel ratio W
Consider feedback in which the input u (k) is corrected with / x hat (k) and x (k) is converged to W. When feedback control for each cylinder is considered with a gain α (k) obtained by accumulating the ratio of the air-fuel ratio x hat (k) estimated by the observer and the target air-fuel ratio W in the model shown in FIG. As shown. At this time, input u
Putting (k), the equations 18 and 19 are obtained.

[0060]

[Equation 18]

[0061]

[Formula 19]

From Equation 18, Equation 20 is derived.

[0063]

[Equation 20]

Expression 21 is derived from Expression 19.

[0065]

[Equation 21]

Therefore, u (k) / u (k-4) ≈1, k
→ When ∞, x (k-4) → If x (k-4), then k →
When ∞, x (k) → W should be obtained.

This can be generally expressed as [current output] = [current input] × [target value] / [current output estimated value] × [specific control cycle pre-correction value]. In the above equation, “before a specific control cycle” means, in this case, before four control cycles (TDC), that is, the same cylinder output in a four-cylinder engine. However, when the feedback was actually simulated with this gain, the control was not stable.

Therefore, in order to add a delay to the cumulative calculation of the gain α (k), the value obtained two times before is used as shown in FIG. At this time, it becomes like Formula 22.

[0069]

[Equation 22]

When calculated in the same way as without delay,
And becomes stable.

[0071]

[Equation 23]

This will be described with reference to FIG.
A / F estimated by a particular cylinder (by observer),
The x hat (k) is the result of control using the correction value α (k) for that time. Therefore, when calculating the correction value, it is necessary to recognize how many times the estimated A / F was before and what the gain value was at that time. In that sense, as shown in FIG. 27, the observer output four times before (the last time when paying attention to the first cylinder) is eight times before (two times before).
No. 1 cylinder air-fuel ratio is estimated. In this way, the next control gain is calculated from the control gain eight times before and the result (estimated value) of control using the control gain. It is something that has been taken and converged. The simulation result is shown in FIG. 28 (the upper part of FIG. 28 shows the result without delay, and it can be seen that it is more stable than that. The solid line in the figure shows the case of feedback control, and the broken line shows the feedback. It shows the case of not controlling). FIG. 29 shows a block diagram of this model (a model obtained by adding a feedback control system to the model of FIG. 11). 30 to 34 show simulation results of this model. It can be seen from FIG. 33 that the convergence is significantly improved compared to the PID control.

As a result of further pursuing the delay here, as shown in FIG. 35, it is preferable to use the control gain 12 times before (three times before for the first cylinder). There was found. That is, inside the engine,
It is possible to accurately represent the delay such as reflecting the control result,
As a result of the clarification of the delay amount determined by the number of cylinders and the delay amount determined by the number of combustion cycles, as shown in FIG. 36, the reflection delay of the control result + completion of one combustion stroke + sampling delay + observer estimation + It was found that the gain 12 times before (three times before in the case of the same cylinder) can be used as the cumulative control gain from the delay in gain allocation. For reference, FIGS. 37 to 40 show control gains to be used in the case of three cylinders, five cylinders, six cylinders, and twelve cylinders.

Next, it is confirmed that the ideal input can be input to the feedback control model shown in FIG. 29 to converge the cylinder-by-cylinder air-fuel ratio to the target value. We also consider the effect of the observer weight matrix.

As an input to the feedback control model shown in FIG. 29, the cylinder-by-cylinder air-fuel ratio is [12.0 / 14.7 / 1].
4.7 / 14.7] and virtual noise [12.0 ± 0.2 / 14.7 ± 0.2 / 14.7 ±
0.2 / 14.7 ± 0.2], the calculation results when each cylinder is independently varied are shown in FIGS. 41 and 42 (FIG. 4).
1, (a) to (c) in FIG. 42 are both (a)
Q element: R element = 1: 10 control result, (b) Q element: R element = 1: 1 control result, (c) Q element:
The control result of the element of R = 10: 1 is shown). However, the target value was set to air-fuel ratio = 14.7, and the calculation was performed when the elements of the load matrix of the observer were Q: R = 1:10, 1: 1 and 10: 1. The control errors at that time are shown in Table 4 and Table 5, respectively. As can be seen from FIG. 41, with respect to an ideal input in which the air-fuel ratio of each cylinder is constant, the heavier the convergence load of the observer, the faster the air-fuel ratio of each cylinder can converge to the target value. As shown in FIG. 42, when the air-fuel ratio of each cylinder is not stable, the convergence accuracy is poor because the feedback is delayed.

[0076]

[Table 4]

[0077]

[Table 5]

Note that in FIG. 25 and below, the feedback control is performed by multiplying each cylinder input air-fuel ratio by the control gain for the sake of simulation, but in reality, the gain is calculated as a multiplication term of the fuel injection time pulse Ti as shown in FIG. It will be.

In the above-described embodiment, the air-fuel ratio of each cylinder is accurately estimated from the detection value of one air-fuel ratio sensor arranged in the exhaust system collecting portion, and the air-fuel ratio of each cylinder is set to the target value accordingly. Because it was configured to control deadbeat,
The convergence of control can be significantly improved.

In the above embodiment, an example is shown in which one air-fuel ratio sensor is used to estimate the air-fuel ratio of each cylinder and the target value is controlled. However, the present invention is not limited to this, and the air-fuel ratio for each cylinder is changed. The present invention can also be applied to a case where a dead ratio control is performed to a target value after a fuel ratio sensor is provided to detect the air-fuel ratio of each cylinder.

[0081]

In the air-fuel ratio control method for an internal combustion engine according to the first aspect of the present invention, the air-fuel mixture input from the output of a single air-fuel ratio sensor arranged in the exhaust system collecting portion of the multi-cylinder internal combustion engine. In controlling the air-fuel ratio to a target value by detecting the air-fuel ratio of the above, the output value of the sensor is considered to be a weighted average value obtained by multiplying the combustion history of each cylinder by a predetermined weight to determine the behavior of the exhaust system. The model to be described is constructed, the state equation and the output equation with the air-fuel ratio of each cylinder as an internal state variable are obtained, the output is obtained by constructing an observer for observing the internal state, and the output of the observer of each cylinder is obtained. The air-fuel ratio is estimated, the ratio between the estimated value and the target value is obtained, and the product of the ratio and the correction value before the specific control cycle is used as the current correction value to immediately converge the air-fuel ratio of each cylinder to the target value. To control dead beat like Having Ranaru as configuration, without preparing the air-fuel ratio sensor separately for each cylinder, with the air-fuel ratio of each cylinder can be accurately detected, it can be immediately converged to the target value.

In the air-fuel ratio control method for an internal combustion engine according to claim 2, the air-fuel ratio of each cylinder of the multi-cylinder internal combustion engine is detected, and the ratio between the detected value and the target value is obtained, and the ratio is calculated. Is multiplied by the correction value before the specific control cycle is used as the correction value this time, and the dead beat control is performed so that the air-fuel ratio of each cylinder immediately converges to the target value. Can be immediately converged to the target value.

In the air-fuel ratio control method for the internal combustion engine according to the third aspect, the specific control cycle is a number corresponding to an integral multiple of the number of cylinders.
The air-fuel ratio of each cylinder can be immediately and more accurately converged to the target value.

[Brief description of drawings]

FIG. 1 is an overall schematic view of an air-fuel ratio detection / control device for an internal combustion engine that realizes a method according to the present invention.

FIG. 2 is a block diagram showing a configuration of a control unit in FIG.

FIG. 3 is a simulation result showing a response delay of the air-fuel ratio sensor in the case where the amount of supplied fuel is changed stepwise while the amount of intake air is constant in a single-cylinder internal combustion engine, data representing the measured value of the LAF sensor output in the same case. Is.

FIG. 4 is a block diagram showing an example in which a detection operation of an air-fuel ratio sensor is modeled.

5 is a model in which the model shown in FIG. 4 is discretized with a period ΔT.

FIG. 6 is a block diagram showing a true air-fuel ratio estimator modeling the detection behavior of the air-fuel ratio sensor according to the present invention.

FIG. 7 is a graph showing an actual measurement value of an LAF sensor output and an estimated value of a true air-fuel ratio obtained by multiplying it by an inverse transfer function.

FIG. 8 is a block diagram showing a model showing the behavior of the exhaust system of the internal combustion engine used in the present invention.

9 is a diagram showing a model shown in FIG. 8 in which a three-cylinder air-fuel ratio and a one-cylinder air-fuel ratio of a four-cylinder internal combustion engine are set to 14.7 and 1 respectively.
It is data showing the case where the fuel is supplied at 2.0.

10 is data showing the air-fuel ratio of the collecting portion of the model of FIG. 8 when the input shown in FIG. 9 is given.

11 is a graph comparing the air-fuel ratio of the collecting portion of the model of FIG. 8 when the input shown in FIG. 9 is given, without correcting the response delay of the LAF sensor, with the measured value of the LAF sensor output in the same case. It is a graph to do.

FIG. 12 is a block diagram showing a configuration of a general observer.

FIG. 13 is a block diagram showing a configuration of an observer used in the present invention.

14 is a block diagram showing a simulation model in which the model shown in FIG. 8 and the observer shown in FIG. 13 are combined.

FIG. 15 is a graph showing a four-cylinder internal combustion engine using the model of FIG.
It is data which shows the simulation result at the time of supplying fuel to 2.0.

16 is data showing simulation results when fuel is supplied with virtual noise included in the example of FIG. 15. FIG.

17 is an explanatory diagram showing estimated deviations in time series for the case shown in FIG.

18 is an explanatory diagram showing an observer output when an actual air-fuel ratio sensor output is input to an air-fuel ratio estimator shown in FIG. 6 and an actual assembly air-fuel ratio is input.

FIG. 19 is a block diagram showing an example in which PID control is performed for each cylinder based on a value obtained from the observer output shown in FIG. 13 so that the air-fuel ratio becomes a target value.

FIG. 20 shows an example of response characteristics of the output air-fuel ratio when the input air-fuel ratio is constant, in the PID control simulation data shown in FIG. 19.

FIG. 21 shows an example of response characteristics of control gain when the input air-fuel ratio is constant, in the PID control simulation data shown in FIG.

FIG. 22 is a simulation data of the PID control shown in FIG. 19, showing an example of the characteristic when the input air-fuel ratio is changed.

23 is an example of the response characteristics of the output air-fuel ratio when the input air-fuel ratio is changed as shown in FIG. 19 with the simulation data of the PID control shown in FIG.

FIG. 24 is a simulation data of the PID control shown in FIG. 16, showing an example of the response characteristic of the control gain when the input air-fuel ratio is changed as shown in FIG.

FIG. 25 is a block diagram of dead beat control according to the present invention.

FIG. 26 is a block diagram similar to FIG. 22, showing an example of control including a delay.

FIG. 27 is an explanatory timing chart showing the reason for determining the control gain of FIG. 23.

28 is data showing contrasting simulation results of the control shown in FIGS. 25 and 26. FIG.

FIG. 29 is a block diagram showing a model used in the control of FIGS. 25 to 26.

FIG. 30 shows simulation data of the model shown in FIG.
The response characteristic of the output air-fuel ratio when the input air-fuel ratio is constant is shown.

FIG. 31 shows simulation data of the model shown in FIG.
The response characteristic of the control gain when the input air-fuel ratio is constant is shown.

FIG. 32 shows simulation data of the model shown in FIG.
The characteristics when the input air-fuel ratio is changed are shown.

FIG. 33 shows simulation data of the model shown in FIG.
32 shows the response characteristics of the output air-fuel ratio when the input air-fuel ratio is changed as shown in FIG.

FIG. 34 shows simulation data of the model shown in FIG.
The response characteristic of the control gain when the input air-fuel ratio is changed as shown in FIG. 32 is shown.

FIG. 35 is a block diagram showing an example similar to the example shown in FIG. 26 and further including another delay.

FIG. 36 is an explanatory timing chart showing the reason for determining the control gain of FIG. 35.

FIG. 37 is an explanatory timing chart showing a control gain in the case of three cylinders, which is an example similar to FIG. 35.

38 is an explanatory timing chart showing a control gain in the case of 5 cylinders, which is an example similar to FIG. 37. FIG.

FIG. 39 is an explanatory timing chart showing a control gain in the case of 6 cylinders, which is an example similar to FIG. 37.

FIG. 40 is an explanatory timing chart showing a control gain in the case of 12 cylinders, which is an example similar to FIG. 37.

41 is data showing a simulation result when an ideal input is input to the model shown in FIG. 29. FIG.

FIG. 42 is data showing simulation results when an input in which virtual noise is added to the model shown in FIG. 29 is input.

[Explanation of symbols]

 10 Internal Combustion Engine 18 Intake Manifold 20 Injector 22 Exhaust Pipe 40 Air-Fuel Ratio Sensor 42 Control Unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Shusuke Akasaki 1-4-1 Chuo, Wako-shi, Saitama Inside Honda R & D Co., Ltd. (72) Inventor Isao Komoritani 4-1-1 Chuo, Wako, Saitama Stock Company Honda Technical Research Institute

Claims (3)

[Claims] 1. An air-fuel ratio is controlled to a target value by detecting the air-fuel ratio of an air-fuel mixture input from the output of a single air-fuel ratio sensor arranged in the exhaust system collecting portion of a multi-cylinder internal combustion engine, a. A model describing the behavior of the exhaust system is constructed by regarding the output value of the sensor as a weighted average value obtained by multiplying the combustion history of each cylinder by a predetermined weight, and the air-fuel ratio of each cylinder is used as an internal state variable. Determine the state equation and the output equation, b. Constructing an observer for observing the internal state and obtaining its output, c. Estimating the air-fuel ratio of each cylinder from the observer output, d. The dead beat control is performed so as to obtain the ratio between the estimated value and the target value, and multiply the ratio by the correction value before the specific control cycle as the current correction value to immediately converge the air-fuel ratio of each cylinder to the target value. An air-fuel ratio control method for an internal combustion engine comprising: 2. A. Detecting the air-fuel ratio of each cylinder of the multi-cylinder internal combustion engine, b. The dead beat control is performed so as to obtain the ratio between the detected value and the target value, and multiply the ratio by the correction value before the specific control cycle as the current correction value to immediately converge the air-fuel ratio of each cylinder to the target value. An air-fuel ratio control method for an internal combustion engine comprising: 3. The specific control cycle is a number corresponding to an integral multiple of the number of cylinders.
An air-fuel ratio control method for an internal combustion engine according to the above item.