JP2526250B2 - Fuel control device for internal combustion engine - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fuel control device for an internal combustion engine, which particularly performs learning control.

(Prior Art) A device in which a fuel control device is provided with a learning function has been proposed (see JP-A-55-96339). This is applied to the air-fuel ratio control for the L-Jetronic fuel injection engine, and the basic pulse width Tp (= K × Qa / N, where K is a constant, Qa is the intake air amount, and N is the number of revolutions. ) Is corrected by the feedback correction coefficient α and the learning correction coefficient KBLRC, the injection pulse width Ti is calculated by the formula Ti = Tp × COEF × α × KBLRC + Ts. Here, Ti is a pulse width of the injection valve drive corresponding to the fuel injection amount required for one ignition cycle, and Tp is a pulse width corresponding to the injection amount for obtaining a constant basic air-fuel ratio. The theoretical air-fuel ratio is adopted as the basic air-fuel ratio, but the basic air-fuel ratio is only a set value, and in reality there is a deviation between the two air-fuel ratios. This deviation is eliminated by α calculated based on the deviation of both air-fuel ratios. COEF is the sum of various correction coefficients (for example, the mixing ratio correction coefficient KMR) for improving the operating condition of the characteristic, and Ts is the invalid pulse width for the voltage correction.

On the other hand, the learning area in which KBLRC (≦ 1) is stored has a large number of small areas (hereinafter referred to as “areas”) with the representative value of the operation variable (the basic pulse width Tp and the rotation speed N) as the coordinate axis. The learning value is updated (learned) individually for each area. For example, learning is performed when Tp and N are in the same area and a predetermined condition (for example, feedback is sampled for several cycles during feedback control) is satisfied (when the learning condition is satisfied). In this case, the learning correction coefficient is the value LMD obtained from the air-fuel ratio sensor (the maximum and minimum intermediate value of α during sampling for several cycles) and the learning correction coefficient that has been in the area up to now (KBLRC (old)). It is a value calculated by a mathematical expression that uses and as variables, and the calculated value (learning correction coefficient KBLRC
(New)) is stored again in the same area.

(Problems to be solved by the invention) By the way, according to such control, deviation from the theoretical air-fuel ratio (target air-fuel ratio) is caused even if the fuel supply means such as the fuel injection valve has variations or changes over time. Is compensated by feedback control and learning control.

However, in such control, if there are large variations in the injection characteristics of the fuel injection valve, that is, ineffective time and flow rate characteristics (corresponding to the above-mentioned Ts and K), the stored learning value KBLRC is stored.
Is a value that is greatly distorted depending on the learning area, that is, the operation area, and therefore, it is not possible to accurately respond to variations in the injection valve depending on the selection of the learning area, and there is a problem that a very good control accuracy cannot be obtained. there were. Also,
If the learning value is distorted in this way, feedback control will be added if the operating conditions at the time of reading the learning value and when the fuel is supplied using the learning value are significantly different, for example, during transient operation. However, there was a problem that the drivability of the engine and the exhaust emission were deteriorated.

An object of the present invention is to provide a fuel control device that solves such problems.

(Means for Solving the Problems) As shown in FIG. 1, the present invention comprises means a for detecting the intake air amount of the engine, means b for detecting the number of revolutions of the engine, and
A fuel supply amount calculation means d for calculating the fuel supply amount from the fuel supply means c based on these operating conditions, and means e for detecting the air-fuel ratio of the air-fuel mixture supplied to the engine.
And an air-fuel ratio correction means f for feedback-correcting the fuel supply amount from the fuel supply means c based on the difference between the detected air-fuel ratio and the target air-fuel ratio, the air-fuel ratio correction means The first storage means g for storing the correction amount by f, the means h for calculating the correction amount of the ineffective time of the fuel supply means c from the stored value of the first storage means g in the low load region, and the correction amount. Second storage means i for storing
And a means j for calculating a correction amount of the flow rate characteristic value of the fuel supply means c from the stored value in the high load region of the first storage means g,
A third storage means k for storing this correction amount and a fuel supply amount correction means 1 for correcting the fuel supply amount from the fuel supply means c by the stored values of the second and third storage means i, k. Set up.

(Operation) With respect to variations and changes with time of the fuel supply means c, the correction amount can be obtained from the feedback correction amount by the air-fuel ratio correction means f, and the responsiveness of the fuel supply means c, that is, the variation, can be obtained. The ineffective time (depending on the drive voltage) changes, but this change has a large value in the low load region. Further, the variation in the flow rate characteristic of the fuel supply means c becomes similar between the high load region and the low load region.

Therefore, from the feedback correction amount in the low load region and the feedback correction amount in the high load region, the correction amount of the dead time of the fuel supply means c and the correction amount of the flow rate characteristic value are separately calculated and learned, and these learned values are obtained. Since the fuel supply amount from the fuel supply means c is corrected by the (memorized value), even if the fuel supply means c varies, each learned value does not take a value greatly distorted depending on the operating region, and the supply space It is possible to reduce the difference between the fuel ratio and the target air-fuel ratio, so that the fuel supply can be controlled accurately and responsively even during steady operation or transient operation, and even during open control without feedback control. Accurate fuel supply control is possible.

(Embodiment) FIG. 2 shows a mechanical configuration in which the present invention is applied to an L-Jetronic fuel injection engine. In the figure, a sensor (for example, a flap type air flow meter) that detects the intake air amount Qa, a sensor (crank angle sensor) 2 that detects a unit angle and a reference position of a crank angle, and a sensor that detects an actual air-fuel ratio ( An air-fuel ratio sensor) 3 is provided in each part of the engine. The rotation speed N is obtained by counting the unit angle signal of the crank angle input during a predetermined time. Further, 4 is a water temperature sensor, 5 is a throttle valve opening sensor, 6 is a knock sensor, 7 is a battery, 8 is a vehicle speed sensor, and 9 is a key switch.

All signals from these sensors are input to the control unit 10, and the unit 10 controls the fuel injection amount of the fuel injection valve 11 based on various operating variables.

FIG. 3 is a block diagram when the control unit 10 is composed of a microcomputer. That is, interface (I / O) 12, CPU13, ROM14, RAM1
5.Non-volatile RAM that can retain stored information even when the power is turned off
In addition to the (BURAM) 16, an AD converter (ADC) 17 for converting an analog signal of various signals into a digital signal is attached.

Next, the control contents executed by the control unit 10 will be described based on the flowcharts of FIGS.

FIG. 4 is a routine for calculating the fuel injection amount (injection pulse width) Ti. The intake air amount Qa, the engine speed N, and the injection constant KCON.
From ST (preset) and the correction amount KLCD1 of the flow rate characteristic value of the fuel injection valve 11, the basic injection amount (basic pulse width) Tp is calculated by the following equation (1) (steps 40 to 44), Tp = (KCONST x KLCD1) x (Qa / N) (1) This Tp, the invalid time (invalid pulse width) Ts due to the drive voltage of the fuel injection valve 11, and the correction amount KLCD3 for the invalid time.
And the total COEF of various correction factors, the reference correction amount KLCD2,
From the feedback correction coefficient α by the air-fuel ratio sensor 3, the fuel injection amount Ti is calculated by the following equation (2) (step
45-53).

Ti = Tp × COEF × KLCD2 × α + (Ts + KLCD3) (2) Here, each correction amount KLCD1, KLCD2, KLCD3 is a value learned separately from the learning value KLCD of the correction amount by the feedback correction coefficient α (described later). .

The various correction coefficients COEF are the sum of the mixture ratio correction coefficient (high load increase correction coefficient) KMR, water temperature correction coefficient KTW, etc. These are the basic injection amount Tp, engine speed N, engine cooling water temperature as shown in FIG. It is read from a map value set in advance according to Tw, etc., and in this case, COEF is added with a variable KTRM that determines whether it is in the feedback control range, and COEF = KTRM (KMR = 0) is selected in the feedback control range. (Steps 60-66). However, KTRM = 0 is taken outside the feedback control region, and at this time, the feedback correction coefficient α in the routine of FIG. 4 is set to 1.0.

Further, as shown in FIG. 6, the ineffective time Ts of the fuel injection valve 11 is Ts at a predetermined voltage (for example, 14V), the drive voltage VB,
It is calculated from the proportionality constant DTS (steps 70 to 73).

Next, the learning value KLCD of the correction amount by the feedback correction coefficient α and the respective correction amounts KLCD1 to KLCD3 separately learned from the learning value KLCD are calculated. The learning value KLCD is the correction amount KLCD1 (1 data). Is equal to the multiplication value of KLCD2 (map value) and is stored in a learning map divided into the same learning area as KLCD2 in this case, and the learning value KLCD is updated when feedback control is performed.

That is, the learning value KLCD is calculated from the correction amounts KLCD1 and KLCD2 by the calculation routine of FIG. 7, while the learning area (Qa / N, N) when KLCD2 is read is stored (step 80).
~ 83), if feedback control is performed in the same area (Qa / N, N) at this time (COEF = KTRM), the maximum and minimum of the feedback correction coefficient α that is proportionally integrated (sampled for several cycles) Deviation of the mean value of LMD ΔLMD ＝ LMD−
1.0 is calculated, and the current KLCD is calculated from the ΔLMD, learning gain LRG (map value assigned to ΔLMD), and the previous KLCD _-1 (equal to the value calculated in step 83) by the following equation (3). , The value of the same learning map area is updated (steps 84 to 88).

KLCD = KLCD ₋₁ ＋ LRG × ΔLMD (3) Then, each correction amount separated and learned from this learning value KLCD
The calculation of KLCD1 to KLCD3 is performed in the routine shown in FIG. In addition, A and B in FIG. 8 are reference values for setting the load region.

First, in the predetermined high load range where Qa / N representing the engine load is larger than the reference value A, the routine of FIG.
If the number of learning airs that learned KLCD in 8) is a predetermined number n, each KLCD read from these areas is compared with a reference value of 1.0, and these deviations are in the same positive or negative direction. Sometimes, the absolute value of the deviation | KLCD−1.0 |
The KLCD that minimizes is selected as KLCD (k). (100
~ 105). Then, this KLCD (k) and learning gain GKC (KL
A correction amount KLCD1 (correction amount of the flow characteristic value of the fuel injection valve 11) is calculated from the map value assigned to CD (k) -1.0) by the following equation (4) and stored (steps 114, 11).
6).

KLCD1 = (KLCD (k) −1.0) × GTC + 1.0 (4) Here, if there is variation in the flow rate characteristics of the fuel injection valve 11, the flow rate tends to either increase or decrease. , KLCD can be learned from the correction amount KLCD1 when the deviation from the reference value 1.0 of the KLCD is in the same direction, and since KLCD1 is calculated from the minimum deviation between the KLCD and the reference value 1.0, the KLCD1 with less error is obtained. To be

On the other hand, in a predetermined low load area where Qa / N is smaller than the reference values A and B, if the number of learning areas where KLCD is learned in this low load area is a predetermined number n, each KLCD read from these areas Compared with KLCD (k) selected in step 105, when these deviations are in the same direction of positive value or negative value, respectively, the absolute value | KLCD-KLCD (k) | of the deviation is the minimum KLCD. Is selected as KLCD (s) (steps 101, 10)
6-110). And the difference between KLCD (k) and KLCD (s) ΔKLCD
Correction amount KLCD the learning gain GTS (map value assigned relative DerutaKLCD) reading, from the GTS and the previous KLCD3 _-1 from
3 (correction amount of the invalid time of the fuel injection valve 11) is calculated and stored (steps 111 to 113, 115).

Here, KLCD (s) obtained from KLCD in the low load range is KLCD.
Since it includes a correction amount of 1, and KLCD (k) obtained from KLCD in the high load range takes an approximate value in the high load range and the low load range, KLCD (s) and KLCD (k) Difference from KLCD3
(1 data) can be learned.

If the conditions of steps 101 and 106 are not satisfied, learning is not performed and the previous KLCD1 and KLCD3 are read (steps 117 and 118).

And the calculation of the reference correction amount KLCD2, which is KLC
The correction amount H3KLCD of KLCD is obtained from D3, and the addition of H3KLCD to KLCD is required to be divided by KLCD1 and stored (steps 119 to 121). This H3K LCD is shown in Figs.
It is obtained by multiplying the map values L and K set for the KLCD3 and the load Qa / N as shown in the figure.

This KLCD2 is obtained by inversely correcting KLCD by KLCD3, that is, by separating the correction amount for KLCD3 in the low load region from KLCD2, and therefore KLCD2 stored in each area of the learning map regardless of the progress of learning. Is a value with less distortion.

When the conditions of steps 102, 104, 107 and 109 are not satisfied, H3KLCD = 0 is set (step 122) and KLCD2 is not updated.

Then, the fuel injection amount Ti is calculated using the correction amounts KLCD1 to KLCD3 in the routine of FIG. 4, and is output as a drive signal for the fuel injection valve 11.

With this configuration, the learning value KL of the feedback correction amount by the air-fuel ratio sensor 3 that detects the actual air-fuel ratio of the engine
From the CD, correction for variations in the fuel injection valve 11 and changes over time, that is, the correction amount KLCD1, of the flow rate characteristic value of the fuel injection valve 11,
The correction amount KLCD3 and the reference correction amount KLCD2 of the invalid time of the fuel injection valve 11 are obtained separately, and these learned values KLCD1 to KL are obtained.
The fuel injection amount Ti is corrected by CD3.

Therefore, even if the fuel injection valve 11 has large variations, the correction amounts KLCD1 to KLCD3 can bring the fuel injection amount Ti close to the required injection amount, that is, the correction amounts KLCD1 to KLCD3.
As a result, the KLCD approaches the reference value 1.0 as shown in FIG. 11, and therefore the difference between the actual air-fuel ratio and the target air-fuel ratio becomes small, so feedback control by the air-fuel ratio sensor 3 can be performed with good response. At the same time, the fuel supply can be accurately controlled even during the open control in which the feedback control is not performed.

In addition, since each correction amount KLCD1 to KLCD3 is learned separately, the learning value does not take a value that is distorted by the operating region, so feedback control is not delayed even during transient operation, and fuel supply High control accuracy can be obtained, and as a result, good engine operating performance and exhaust performance can be secured regardless of steady operation, transient operation, and open control.

(Effect of the Invention) As described above, according to the present invention, the flow rate characteristic value of the fuel supply unit and the correction amount of the dead time are separated and learned from the correction amount in the high load region and the low load region by the air-fuel ratio correction unit. ,
Since the fuel supply amount is corrected and controlled by this learned value, the difference from the supply amount required for the engine can be reduced even if there are large variations in the fuel supply means, changes over time, and the like. The feedback control by the air-fuel ratio correction means can be accurately performed with good response regardless of the transient operation, and the fuel supply can be accurately controlled even during the open control without feedback control, which improves the operating performance and exhaust performance of the engine. be able to.

[Brief description of drawings]

FIG. 1 is a configuration diagram of the present invention, FIGS. 2 and 3 are configuration diagrams showing an embodiment of the present invention and configuration diagrams of a control unit, FIGS. 4 to 8 are flow charts showing control contents, and FIG.
Fig. 10 and Fig. 10 are characteristic diagrams of the data used in the calculation, and Fig. 11 is a graph showing the change in the feedback correction amount as the learning progresses. 1 ... Intake air amount sensor, 2 ... Crank angle sensor, 3
...... Air-fuel ratio sensor, 10 ...... Control unit, 11 ...
… Fuel injection valve.

 ─────────────────────────────────────────────────── ─── Continuation of front page (56) Reference JP-A-58-48739 (JP, A) JP-A-61-129443 (JP, A) JP-A-62-87645 (JP, A)

Claims (1)

(57) [Claims] 1. A means for detecting an intake air amount of an engine, a means for detecting an engine speed, and a fuel supply amount calculating means for calculating a fuel supply amount from a fuel supply means based on these operating conditions. A means for detecting the air-fuel ratio of the air-fuel mixture supplied to the engine, and an air-fuel ratio correction means for feedback-correcting the fuel supply amount from the fuel supply means based on the difference between the detected air-fuel ratio and the target air-fuel ratio. In a fuel control device for an internal combustion engine, comprising: first storage means for storing a correction amount by the air-fuel ratio correction means, and correction of a dead time of the fuel supply means from a stored value in a low load region of the first storage means. Means for calculating the amount, second storage means for storing this correction amount, means for calculating the correction amount of the flow rate characteristic value of the fuel supply means from the stored value of the first storage means in the high load region, Store this correction amount A fuel control device for an internal combustion engine, comprising: a third storage means; and a fuel supply amount correction means for correcting the fuel supply amount from the fuel supply means based on the stored values in the second and third storage means. .