JPS5838354A - System for controlling air-fuel ratio of internal- combustion engine - Google Patents

Abstract

PURPOSE:To improve the fuel consumption of an engine remarkably by a method wherein the air-fuel ratio of a mixture supplied to the engine is controlled so that the fluctuation of combustion of the engine becomes to a constant value and the air-fuel ratio is kept at the best point of fuel consumption before the zone of misfiring at all times. CONSTITUTION:A fluctuating value Tn of a torque is read in steps 11, 12, 13 and a value of Q/N is read in the step 14 while a revolving nuber NE of the engine is read in the step 15. In the step 16, the peak value TA of the fluctuating value of the torque when the engine is operated at the best point of the fuel consumptions under respective conditions by the value of NE and Q/N already read is read out of a map in a memory. In the step 17, a comparison between the peak value TA of the fluctuating value of the torque and the fluctuating value Tn of the torque read in the step 13 is decided and the air-fuel ratio is corrected to the rich side or the lean side thereof in the next step.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an air-fuel ratio control device for an internal combustion engine, and more particularly to an air-fuel ratio control device for controlling the air-fuel ratio at a lean air-fuel ratio at the operational limit of the internal combustion engine.

The components of the exhaust gas emitted by combustion in the internal combustion engine and the torque of the internal combustion engine are determined by the set air-fuel ratio (A/T) of the carburetor that supplies a mixture of fuel and air to the internal combustion engine, as shown in Figure 1. There is a close relationship. Therefore, there are various methods for purifying exhaust gas from automobiles, but from the standpoint of air-fuel ratio, they can be broadly divided into three types: 1) Silicon oxide N shown in A in Figure 11
The air-fuel ratio is set to a rich mixture region with low oxygen emissions, a catalytic purification device is installed in the exhaust pipe, and additional air is supplied to the secondary air supply device to remove unburned components from oxidation. Method for purifying carbon CO and hydrocarbon HC, 2) C in Figure 1
3) A three-way catalyst is installed in the exhaust pipe, and the air-fuel ratio is set near the stoichiometric air-fuel ratio shown in Figure 1 OB. However, a method has been proposed in which 9 Co, HC, and NOx are simultaneously purified using a three-way catalyst.

Recently, in addition to cleaning exhaust gas, there has also been a demand for reducing the fuel consumption rate of engines from the standpoint of resource conservation. As shown in Figure 1, it is advantageous to operate an internal combustion engine in a lean mixture range to simultaneously achieve exhaust gas purification and fuel consumption reduction, but in a lean mixture range, problems such as misfires occur. Therefore, when considering the variation and deterioration of the engine and other accessories (carburizer system), the #1 engine cannot be operated in the lean mixture range where the misfire limit is near, and the misfire limit is reached. As shown in Figure 1, the minimum fuel consumption rate is achieved when the engine is operated at a lean air-fuel ratio just before the misfire region. Combustion fluctuations (torque fluctuations in the present invention) are related to ~, and the closer the misfire region is, the sharper the combustion fluctuations become. The present invention was made in view of the above, and By determining the engine combustion fluctuations and controlling the air-fuel ratio supplied to the engine so that the engine combustion fluctuations are at a certain constant value, you can always prevent misfires and control the air-fuel ratio to the best fuel efficiency point. The primary objective is to achieve a significant improvement in fuel efficiency. The constant value is a value that varies depending on the engine conditions, and in the present invention, the combustion fluctuation control target value is calculated based on the detection signal of the engine sensor, and is set as the optimum value suitable for the engine operating condition. Therefore, this value is a value that changes depending on the operating conditions.

Furthermore, in the present invention, a detection signal of combustion fluctuation (in the present invention, torque fluctuation) is inputted to the control circuit through a filter circuit having filter characteristics in a certain frequency band, so that the detection signal caused by combustion fluctuation of the engine is inputted to the control circuit. The second step is to accurately extract combustion fluctuations.
The purpose of

The present invention provides an air-fuel ratio control device for an internal combustion engine, comprising: an engine sensor for detecting the operating state of the internal combustion engine; and a control circuit for controlling the air-fuel ratio supplied to the internal combustion engine based on a detection signal from the engine sensor. , a torque fluctuation detector that detects combustion fluctuations of the internal combustion engine as torque fluctuations, a filter circuit connected to the torque fluctuation detector and having filter characteristics in a frequency band, and an output from the filter circuit as input, and A control target value for combustion fluctuation of the internal combustion engine is calculated based on the detection signal of the engine sensor, this target value is compared with an output value from the filter circuit, and the difference between the target value and the output value is calculated. There is provided an air-fuel ratio control device for an internal combustion engine, characterized in that the control circuit is configured to correct the air-fuel ratio supplied to the internal combustion engine accordingly.

An air-fuel ratio control device for an internal combustion engine as a first embodiment of the present invention is shown in FIGS. 2 to 17.

In Fig. 2, an internal combustion engine 1 is a spark ignition type engine for driving a car, and air for combustion is supplied to an air cleaner 2゜air 70-.
Meter 3. Via the intake conduit 4 and the intake valve 5, it is sucked into the combustion engine 6 of the internal combustion engine 1. The suction conduit 4 is provided with a throttle valve 7 that can be operated arbitrarily by the driver.Fuel is injected and supplied from an electromagnetic fuel injection valve 8 installed in the suction conduit 4 to the intake valve 5. . A mixture of fuel and air is combustion 1i! 6 and is emitted into the atmosphere via an exhaust valve 9 and an exhaust pipe 10. Fuel Control Ennit 11 This system calculates the amount of fuel supplied to the engine 1 according to the operating state of the engine 1, drives the electromagnetic fuel injection valve 8, and controls the amount of fuel supplied to the engine 1. Air flow meter that detects air volume3. Ignition coil 12. A detection signal from a torque detector 13 that detects torque fluctuations of the engine 1 is input. In this embodiment, the air 70-meter 3 signal is used as the intake air amount to the engine 1, but in addition to the air 70-meter 3, the intake pipe negative pressure generated downstream of the throttle valve 7 of the engine 1 and the engine The amount of intake air may be determined from the rotational speed, or the rotational speed may be determined by detecting a rotational signal from a ring gear or distributor that rotates in synchronization with the engine's 10 rotations. As shown in Fig. 2, the torque detector 13 is attached to a mount 135 that supports the engine by a gel bolt, and detects vibrations centered on the crankshaft in the lean band of the engine using a piezo element, etc. This is to obtain an analog signal proportional to the target torque fluctuation.
In this example, one engine is placed at a symmetrical position.
A total of 2 pieces are arranged. Of course, the present invention can be sufficiently achieved even with a configuration in which only one is disposed for one engine. Torque detector 13 is pressure sensor 131
．． Rubber mount 133 and column mount copper 134
Pressure sensor 131° rubber mount cover 134. The rubber mounts 133 are mounted one on top of the other in this order. As the pressure sensor 131, for example, a commercially available pressure detector having piezo elements arranged in four directions and capable of detecting vibrations (mechanical torque fluctuations) in all directions is used.

Next, the fuel control unit 11 shown in FIG. 3 will be explained. The amplifier is composed of buffers and amplifiers from 20 manufacturers, but since this is well known, the details will be omitted. ZZ band/mother soot filter 30 Outputs only the output of the analog signal from the amplifier 20 at a frequency of IHi to several Hz. It has a multi-configuration including an oscillation circuit using a vibrator and a counter that divides the frequency of this oscillation circuit, and since it is well known, the circuit configuration and detailed explanation will be omitted.

The timing pulse generation circuit 50 is a circuit that generates a recess F signal to the peak hold circuit 60 and an interrupt signal to the correction calculation circuit 100 based on the clock signal from the clock circuit 40 . Its internal circuit is shown in FIG. In FIG. 4, the input terminals 510 and 511 are connected to H? circuit 40
The 0 input terminal 510, into which the 2 Hz and 5 kHz clocks are respectively input, is connected to the reset terminal R of the counter with divider 501, and the input terminal 511 is connected to the ft4
It is connected to the clock terminal CL of the counter with reader 501. The counter 501 with a divider is an IC manufactured by RCA.
D 4017 is used for the pass, and its output Q1 is connected to terminal 51.
The O outputs Q5 and Q8, which are used as signals for interrupt calculation of the correction calculation circuit 100, are connected to the set terminal S and reset terminal R of the R-8 flip-flop 502 through sb, and the output Q9 is clock enable. It is connected to terminal CEK.

R-87 lip flop f502 is RCA IC CD4
013 is used, and its output Q is connected to the peak hold circuit 60 via a terminal 513.

To explain the operation of the timing pulse generation circuit 50 with the above configuration, the 2H% pulse of Figure 6) is input to the reset terminal of the counter with divider 501, and when the pulse changes from "1" to "0", it starts counting. Start. A clock with a frequency of 5 kHz is input to the clock input of the counter 501. Therefore, when the first pulse arrives, a pulse is output to the output Q1. When the ninth pulse comes, the output Q9 becomes r I J and the clock enable terminal becomes "1", so the input of the clock is stopped until the next reset. Therefore, the output Q1 is
) is output as shown below. The output /4 pulse becomes -f pulse via the terminal 512 when the interrupt calculation of the correction calculation circuit 100 is started. Outputs Q5 and Q8 are R-8
7. Set and reset the defroshie 7"502 and set the R
Figure 6 (
A pulse of q is output.

#The pulse is sent to the peak hold circuit 60 via the terminal 513.
The signal is inputted into the peak hold circuit and has a pulse width of about 600 microseconds, which becomes a reset signal for the peak hold circuit.

FIG. 5 shows the internal circuit of the peak hold circuit 60. In FIG. 5, a diode 601 (D positive electrode and 61
The negative electrode of the diode 601 is connected to the output of the bandpass filter 30, and the negative electrode of the diode 601 is connected to one end of the resistor 602. The other end of the resistor 602 is connected to the positive terminal of the welfare capacitor 603, the non-inverting input of the buffer amplifier 606, and the resistor 604.
It is connected to. The negative electrode of the capacitor 603 is grounded.

The other end of the resistor 604 is connected to one end of an analog switch 605. The other end of the analog switch 605 is grounded, and the control terminal is connected to the signal shown in FIG. 6(C) of the timing pulse generating circuit 5o. The buffer amplifier 6060 inverting input is connected to the output. The positive terminal of the diode 611 is connected to one end of a resistor 612. The other end of the resistor 612 is connected to the negative terminal of the capacitor 613, the non-inverting input of the buffer amplifier 616, and the resistor 614.The positive terminal of the capacitor 613 is grounded.

The other end of the resistor 614 is connected to one end of an analog switch 615. The other end of the analog switch 615 is grounded, and the control terminal is connected to the signal of the timing pulse generating circuit 50 shown in FIG. 6(c). The inverting input of buffer amplifier 16616 is connected to the output. The output of the buffer amplifier 606 is connected to one end of the resistor 622, and the other end is connected to the non-inverting input of the buffer amplifier 625 and the resistor 622.
It is connected to 21. The other end of the resistor 621 is grounded. The output of the buffer amplifier 616 is connected to one end of the resistor 623, and the other end is connected to the inverting input of the buffer amplifier 625. The output of the buffer amplifier 625 is outputted to the analog-r isotal (A-D) conversion circuit 70 via an output terminal 633 and connected to one end of a resistor 624. The other end of resistor 624 is connected to the inverting input of buffer amplifier 6250.

To explain the operation of the peak hold circuit 60 with the above configuration, the control inputs of the analog switches 605 and 615 are connected to the tie-building pulse generating circuit 50 as shown in FIG.
When a pulse of
, 613 is discharged through resistors 604, 614 of low resistance value, and the voltages of capacitors 603, 613 are reset to Ov. Thereafter, when the output waveform shown in FIG. 6 (b) of the band/yas filter 30C comes in from the input terminal 632, the capacitor 603 is charged to a positive voltage through the diode 601 and the resistor 602. The voltage of this capacitor 603 is held at a positive peak value after being reset until the next reset. The capacitor 60
When the voltage No. 3 is outputted through the next buffer amplifier 606 with high input impedance, the waveform shown in FIG. 6(G) is obtained. On the other hand, the output waveform shown in FIG. 6(b) is the input terminal 63.
2, the diode 611. Capacitor 613 is charged to a negative voltage through resistor 612. The negative peak value of this capacitor 6130 voltage is held from the time it is reset until the next reset. When the voltage of this capacitor 613 is outputted via the next buffer amplifier 616 with high input impedance, as shown in FIG. 6 (F5
The waveform will be The output of the buffer amplifier 606 and 61
By taking the difference between the outputs of 6 and 6 in the differential amplifier 625, the difference between the positive peak value and the negative peak value from one reset to the next reset is output from the differential amplifier 625, and its waveform is is the first in Figure 6.

FIG. 7 shows an internal circuit diagram of the A/D conversion circuit 70. In FIG. 7, the input/output control (Ilo) signal d of the correction calculation circuit 100 is directly input to the NAND gate data 3, and the inverter 705
is inverted and input. The f vice select signal (SEL signal) of the correction calculation circuit 100 is directly NANIII'-
) 703 and ANII"-) 706. A delay circuit is configured by an inverter 707, a resistor 708 and a capacitor 709, and the SEL signal is input to ANDP-) 706 via this delay circuit. Then, ANDr-) 706 has a width of 10 as shown in Figure 6
Outputs a /4 pulse signal of approximately 0 nanoseconds. This pulse signal is input to the successive approximation ff1A-D converter 701CIA-D conversion command terminal CNV. As the MU converter 701, an ADC80 MU G-12 manufactured by Purge Rown Co., Ltd. may be used.

The conversion end terminal of the A-D converter 701 is connected to the busy terminal BUgYK of the mcH correction calculation circuit 100, and the output terminal B1
to B 12a3 are connected to the path line of the correction calculation circuit 100 via the state passive 702. The 3-stepper v7a 702 is, for example, Toshiba IC TC501.
2 should be used. 'The operation of the Moop conversion circuit 70 configured as described above will be explained. When the pulse shown in FIG.
& Deepen the science program. In the program, the pulse shown in FIG. 6 @) is applied to the A-D conversion command terminal (2) of the A-D converter 701 in response to an A-D conversion start command, and the conversion operation is started at the rising edge of this pulse. At the same time K11E
The output signal of the conversion end terminal EOC shown in Fig. 6 (I) is “1”.
” rise to the level. Here, the conversion end terminal IOC is connected to the pizoi terminal BUSY of the device control unit DCU of the correction calculation circuit 100), and the peak hold circuit 60
The completion of the analog signal reading command from the conversion end terminal E
The user is forced to wait until the OC output signal falls to the "0" level, and until this time, both the I10 signal and the SEL signal are "1".
' level. The successive approximation type A-D converter 701 performs the conversion operation while the output signal of the IOC terminal is at the "1" level, and outputs the f-digitized binary r-digital signal from the output terminals B1 to B12. When the moop conversion operation is completed, the output signal of the conversion end terminal EOC becomes "0".
” level, the standby state of the read command of the correction calculation circuit 100 is released, and the analog signal data from the peak hold circuit 60 is read into the correction calculation circuit 100. A circuit is shown. The rotational speed detection circuit 80 is composed of a /4 pulse shaping circuit 80m and a needle count circuit 80b.The pulse of the minus terminal of the coil of the ignition coil 12 is inputted to the O/base pulse shaping circuit 80m through an input terminal 817. The input terminal 817 has a resistor 80
1C) connected to one end. The other end of the resistor 801 is connected to the resistor 802 and the capacitor 803, and the capacitor 80
The other end of 3 is grounded. The other end of the resistor 802 is connected to the anode of a diode 804.
The cathode of resistor 805. It is connected to a capacitor 806, a Zener diode 807, and a resistor 808. Resistance 805. Capacitor 806 and Zener diode 807
The other end of the resistor 808 is connected to the ground of the transistor 818. The terminal of the transistor 818 is grounded, and the collector is input to the resistor 809 and the Schmitt NAND gate 810. . Connect one end of the resistor 809 to NAND P-) 810
+5V electricity @VC is input to the other person's power. The output of the seamit NAND gate 810 is a monostable multivibrator 813 consisting of a core 'f7 811 and a resistor 812.
The output of the monostable multivibrator 813 is input as a trigger/4 pulse to the capacitor 814 and the resistor 81.
The monostable multivibrator 813゜816 is inputted as a trift9 pulse to the monostable multivibrator 816 composed of RCA IC.
D4047 may be used. Thus, from the output of the monostable multivibrator 816, tl! In response to the signal from the ignition coil 12 shown in FIG. 9 (4), a tie i/da/4 pulse signal having a waveform as shown in FIG. 9 (B) is output.

Next, the counting circuit 80b will be explained with reference to FIG.

The binary counter 851 counts and frequency-divides the clock-to-base pulse signal CI input to the clock terminal CL, and uses, for example, a CD4024 manufactured by RCA.

This counter 851 divides the clock pulse signal C1 of approximately 128 kHz as shown in FIG. 9(C) to divide the clock pulse signal C1 into approximately 32 kHz/4 pulses as shown in FIG. 9(D). A signal is output from output terminal Q2. The counter 852 with a depmeter basically counts the clock pulse signal C1 input to the black and red terminals CL.
The output signal of one of the output terminals Q2 to Q4 is not at the "1" level, and the count operation stop terminal K
When the NKrlJ level signal is input, the counting (number of stitches) operation is stopped.

In this embodiment, the output terminal Q4 and the stop terminal are connected, and when the output of the output terminal Q4 reaches the "1" level, a "1" level signal is input to the stop terminal IN, and the counting operation is stopped. do. In this state, the pulse shaping circuit 80
When the tying Δ pulse signal shown in Figure 9 (0) is input from m to the reset terminal 8, the counter 852 is reset and the output of the output terminal Q4 becomes "0" as shown in Figure 9 ■.
level. Then, when time T has elapsed and the signal input to reset terminal R reaches the "0" level, counter 852 starts counting, and output terminals Q2. Q3
Pulse signals are sequentially outputted from each of them as shown in FIG. 9 (6) and (7), respectively. Thereafter, when the output of the output terminal Q4 reaches the "1" level, the counter 852 stops counting again. The output signals of the counters 851, 852 and the pulse shaping circuit 80& are respectively N0Rr-to 85
The Q3 output of the counter 852 is input to the reset terminal R of the counter 855.

That is, by taking the output signal of the pulse shaping circuit 80m shown in Fig. 9) and the Q3 output 0NOR logic of the counter 852 shown in Fig. 9(G), the fi NOR gate 853
From #! The 14th pulse signal shown in Figure 9 (b) is output, and the output signal of this N0Rr-)853 and the 9th
By performing a NOR operation with the output signal of the counter 851 shown in FIG. 855.

Here, the timing pulse signal shown in Figure 9 (0) is "0".
N0Rf-) 853 shown in the building 9 Zuchi below
At time t1, when the output of the counter 855 reaches the rlJ level, the counter 855 stops counting. The outputs of the output terminals Q1 to Q12 of the counter 855 are at time t2.
Shift registers 856 to 858 (for example, RCA CD4)
035) is temporarily held and stored. Next, when the Q3 output of the counter 852 reaches the "1" level at time t3, the counter 855 is reset, and when the Q4 output of the counter 852 reaches the "1" level at time t4, the counter 855 starts counting again. Since the operation of this counter 855 is repeatedly performed in synchronization with the ignition coil 12 outputting the ignition signal, each of the output terminals Q1 to Q4 of the shift registers 856 to 858 outputs the reciprocal of the engine rotational speed N, 1//N. A two-over signal proportional to is output.

The 3-state buffer 860 outputs “1” to the control terminal 861.
While the level signal is applied, the output becomes high impedance, and the output terminal group 859 is connected to the correction calculation circuit 100 via a pass line. The control terminal 861 receives the output signal of the NAND gate 862. NAND l)-) 862 is a device control unit built in the correction calculation circuit 100
The I10 signal and 8KL signal from CU) are input. Then, the output signal of the NAND gate 862 becomes “0”.
” level, shift registers 856 to 858
A two-over signal proportional to the IA of is input to the correction calculation circuit 100.

Next, the intake air amount counting circuit 90 will be explained with reference to FIG. A clock pulse of about 128 k's as shown in the 11th WIJ (B) is input to the input terminal 911,
The signal is input to the NAND gate 902 and the clock terminal CL of the counter 901 with depaging. A time TPO pulse proportional to the amount of intake air per engine revolution (hereinafter referred to as 乍) as shown in FIG. 9
02 and is input to the reset terminal R of the counter 901 with a depmeter. Counter with nuclear depimeme 901 #1 RCA company g
IcocD4017 may be used. Counter 901Fi with divider.

Basically, the output voltage is input to the black V terminal CL.
It counts the pulse signal C1, and output terminals Q2 to Q
When the output signal of one of the output terminals 6 is not at the "1" level and the count operation stop terminal ENKrlJ level signal is input, the counting operation is stopped.

However, in the last example, the output terminal Q6 and the stop terminal EN are connected, and when the output of the output terminal Q6 reaches the "1" level, the stop terminal lNKr1J level signal is input, and the counting operation is stopped. . In this state, the fuel amount calculation circuit 12
When the pulse signal shown in FIG. 11 (4) from 0 is input to the reset terminal R, the counter 901 is reset.
A one-company counting operation is started, and pulse signals are sequentially outputted from the output terminals Q2 and Q4 as shown in FIG. 11 Φ) and @)K, respectively. After that, the output of output terminal Q6 becomes “1”.
'' level, the counter 901 stops counting again. Outputs Q2 and Q4 of the counter 901 are input to the clock terminal CL of shift registers 904 to 906 (eg, IC, CD4035 manufactured by RC Co., Ltd.) and the reset terminal R of the counter 903, respectively. Also NAND
The output of dart 902 is clock terminal C of counter 903.
It is input to L.

Here, the time Tp(
When the DA pulse is input to the input terminal 912, a clock C1 related to time T is input to the counter 903, and the number of clocks during this period is counted. (That is, the time Tp is counted. Then, at time 10, when the pulse reaches the "0" level, the counter 903 stops counting. Next, at time t1, the output Q2 of the counter 901 goes to the "1" level. , the output terminals Q1 to Q12 of the counter 903
The outputs of are temporarily held in shift registers 904 to 906. At time t2, the output Q4 of the counter 901 becomes "
1'' level, the counter 903 is reset and enters a standby state for the next counting operation. Since the operation of the counter 903 is repeated in synchronization with the pulse shown in Figure 11, each of the shift registers 904 to 906 is A two-way signal proportional to dQAK is output from output terminals Q1 to Q4 in synchronization with engine rotation. 3 state buffer 9
The expansion output to which a "1" level signal is applied to the 07 safety control terminal 909 has a high impedance, and the output terminal group 908 is connected to the correction calculation circuit 100 via a pass line. The control terminal 909 has NAMDI)f-
) 910 is input, and the NAND gate 910
I10 from the DCU built in the correction calculation circuit.
A signal and a SEL signal are input. and NAND
When the output signal of the gate 910 becomes the “0” level, the two-over signal proportional to the shift registers 904 to 906 is input to the correction calculation circuit 100. The correction arithmetic circuit (hereinafter referred to as CPU) uses a 100-bit, 12-bit i-microcomplex TLC8-12A made by Toshiba.

#i The circuit and operation state of the microcomputer wing are well known, so they will be omitted.

The digital-to-analog (DA) conversion circuit 110 will be explained using the following KIHz diagram. D-A conversion circuit 110
is Invar 11101, NAMDI'-)11
0S! , shift registers 1103 to 1105 and D
1106 (e.g. Parr Brown DAC8)
0). And I1 of CPU100
0 signal is inverted by inverter 1101 and then NAND
The SEL signal is input to the P-t 1102, and the SEL signal is directly input to the i!
(NAND) 1102. Therefore C
When a command is issued to output the air-fuel ratio correction value Fd calculated by the PU 100 to the DA conversion circuit 110, the I10 signal becomes "
0” level, the SEL signal goes to “1” and becomes 4fiy.
NAND r-) 1102 outputs a "0" level signal. This "0" level signal is applied to each shift register 1.
It is input to clock terminals CL of 103 to 1105. Shift registers 1103 to 1105 are the same as those used in the rotation speed detection circuit 80, and when a "0" level signal is input to the clock terminal CL, they take in the signals applied to the data input terminals D to D4. , outputs the signal from output terminals Q1 to Q4. In this way, the two-pass data signal of the air-fuel ratio correction value Fd is input to the input terminals B1 to B12 of the D-me converter 1106, converted into an analog voltage, and then outputted from the output terminal OUT. An analog voltage proportional to the data signal indicating the air-fuel ratio correction value Fd is output from the output terminal OUT.

Next, the fuel amount calculation circuit 120 will be explained.◎Circuit 1
20 is %Il! No. 1976-67016 A device having the same function as the publicly known 4-cylinder engine electronically controlled fuel injection device (hereinafter referred to as KFI),
The intake air amount signal from the air 70-meter 3 and the ignition signal synchronized with the crank rotation of the engine from the ignition coil 12 are input, and the basic valve opening time Tp (hereinafter abbreviated as the injection valve) of the electromagnetic fuel injection valve (hereinafter abbreviated as injection valve) is The injector opening time is determined by calculating the above-mentioned engine speed (time proportional to the amount of intake air at the engine speed), performing various correction calculations depending on the operating condition of the engine, and determining the injector opening time. 8 to control the amount of fuel supplied to the engine 1 shown in FIG. Here, the air-fuel ratio correction value Fd calculated by the correction calculation circuit 100 and converted into an analog voltage by the D-me conversion circuit 110 is corrected by the same method as other correction calculations for intake air temperature and water temperature trend. ing.

The operation of the CPU 100 with the above configuration will be explained along the flowchart of FIG. 43. When a key switch (not shown) is turned on, the power is turned on and the operation starts. Clear all memory in step 1 (81), rOJKl, then Stella 7
"2 (82) sets the initial value of the air-fuel ratio correction value Fd to 2048 (
(12-bit center value) and set flag C, which is used to determine the magnitude of torque fluctuation, to 1. Stella f3 (83
) to set the master mask so that the CPU accepts interrupt operations, and then step 4 (84) to enter the standby state for interrupt operations. After that, as time passes, the timing/4 pulse generation circuit 5
6 from 0 (interrupt operation is started when the carrier pulse rises from rOJ to "l". When interrupt operation is started, subsequent interrupts are prohibited at Stella 7'l0 (810). Stayft1 ( sxl) in FIG.
At the same time as starting A-D conversion, the IOC terminal output (Fig. 6 (■))/f pulse becomes "1" and the CPU 100
When the BU8Y input becomes "1", the operation is stopped. This is stay7'12 (812)・A-D converter 701
When the A/D conversion is completed, the EOC terminal output α) in FIG. 6 becomes "0" and the CPU 100 resumes calculation. When the calculation is restarted, Stella 7'13 (81B) is A-
The torque fluctuation value Tn output from the D converter 701 is read into the CPU 100. As above Stella 7'll
, 12°13, the torque fluctuation value Tn is uniformly converted and read. In step 14 (814), the intake air amount stitch number circuit 90 counts and the value of 9ψ is read. Stella 7'l! 5 (815), a value 1/NE which is inversely proportional to the engine rotational speed counted by the rotational speed detection circuit 80 is read, and the engine rotational speed NK can be obtained by taking the reciprocal of this value.

Figure 14 is a map showing the peak torque fluctuation value (liTA) when the engine is operated at the best fuel efficiency point under each condition, with dNK and Φ conditions as parameters for the engine operating conditions. This is a map showing the read-only memory (hereinafter referred to as ROM). ).In step 16 (816), the TA stored at the corresponding ROM address is read out by searching for the location of the NE in the map shown in Figure 14 for the read data. Stella f17 (Sl'y) determines the magnitude of TA and Tn read by stage 13, and determines whether the air-fuel ratio under the current engine operating condition is the best fuel efficiency point or rich or rich. In other words, if Tn)TA, the torque fluctuation is large and it is determined that the air-fuel ratio is lean, and in steps 18 to 24 (818 to 24) the air-fuel ratio is rich-corrected according to the difference between the output value of the torque fluctuation and the target value. If T n (TA), the torque fluctuation is small and the air-fuel ratio is determined to be rich, and steps 25 to 31 (82
5 to 31), the air-fuel ratio is corrected to be lean in multiple steps. Step 1B is rich/lean discrimination flag C
1, and keeping in mind that the rich correction calculation was performed, the air-fuel ratio proportional to the difference between the current torque fluctuation value Tn and the target torque fluctuation value Tm is set to 9. Correct. That is, if the difference in torque fluctuation values is large, the air-fuel ratio is corrected to a large extent, and if the difference is small, the air-fuel ratio is corrected to a small value. Stesive 20°21 is the air-fuel ratio ratio,
FM correction value FaC) maximum value is FM. The above-mentioned maximum caution was determined as an air-fuel ratio correction value sufficient to bring the engine from the air-fuel ratio at the boundary between the unstable combustion region and the safe combustion region to the air-fuel ratio in the stable combustion region. Step 22 calculates the current air-fuel ratio correction value Fd by adding the correction value Fa obtained in steps 19 to 21 to the air-fuel ratio correction value Fd obtained in the previous calculation. Stesive 23.24 sets the air-fuel ratio correction value Fl to 1.
The maximum value is regulated by a 2-bit maximum value of 4095. At step f25, the flag C is checked to determine whether the interrupt operation performed previously was a rich correction or a lean correction. Calculate the lean correction value PL with Stella f26 to 28 when Fi (when C-1) is applied to the previous calculation.
J, 15th, which stores in advance in the ROM how many times the lean correction value Fa is to be divided into lean correction up to the target value.
Read out the table in the figure. Stella f27 divides the rich correction value Fa by the number of divisions of lean correction obtained in this way to calculate the lean correction value FL for each time. In step 28, the flag C is set to 0', and when the next interrupt calculation is to continue the lean correction calculation, the lean correction value FL is set to the same value as the previous one. Stella 7'2
Step 9 calculates the current air-fuel ratio correction value Fd by subtracting the lean correction value FL from the previous air-fuel ratio correction value Fd. In steps 30 and 31, the minimum value of Fd is set to 0 so that the air-fuel ratio correction value Fd does not become a negative number due to the calculation of the static 29. Stella 7'32 (832) applies W to the air-fuel ratio correction value Fdt-D-A conversion circuit 110 obtained by the above calculation, allows interrupt in Stesive 33 (833), and calculates Stella 7'34 ( At step 834), the state returns to the Grodaram dormitory row state before the interrupt occurred. With the above configuration and operation, the air-fuel ratio supplied to the engine can be controlled to the best fuel efficiency point based on the detection signal from the torque detector 12. The time chart of the air-fuel ratio correction calculation in the flowchart of FIG. 13 is shown in FIG. 16. 7
This shows the state of lag C, and FIG. 16(C) shows an analog voltage obtained by DA converting the air-fuel ratio correction value y-. When the torque fluctuation signal shown in Figure 16 becomes larger than the torque fluctuation target value T at the best fuel efficiency point, rich correction is performed according to steps 18 to 24 of the flowchart in Figure 13;
If it becomes small, lean correction is performed according to steps 25 to 31. In addition, in this example, the torque fluctuation value at the best fuel efficiency point is, for example, the engine operating condition of 2000 rp shown in Fig. 1.
m, 4 to 9-m, it is 0.1 kl-m, and the relationship between the torque fluctuation value and the detection signal obtained from the peak hold circuit is as shown in Fig. 17, so the judgment level under the engine operating conditions is TA (Stella 7 in the flowchart in Figure 113)
'17) is the output voltage from the peak hold circuit and is 1.5
v. The way to determine the rich correction value Fa when performing rich correction is to correct the fuel consumption once (every 0.5 seconds) from the air-fuel ratio at the boundary between the misfire region and the stable combustion region (air-fuel ratio 20.5 as shown in Figure 1). In order to correct to the air-fuel ratio at the best point and in the stable combustion range (air-fuel ratio 20 in the It diagram), the rich correction value Fa for one time should be 0.5 in terms of air-fuel ratio, and this should be set to the maximum value. It was set to FM. and the air fuel ratios 20.5 and 20.0.
The rich correction value is 0.5 per time based on the difference in torque fluctuation values.
The proportionality constant was determined to be 1 to obtain the following. (When correcting to the rich side, it is important to quickly correct the air-fuel ratio to the stable combustion range in one correction to prevent engine misfires. FLFi The previous rich correction value F1 was calculated by equally dividing it into a number of divisions suitable for this value.The number of divisions reaches the target value in 20 seconds even when the air-fuel ratio changes by 05, which is the time when the maximum air-fuel ratio changes. Thus, D=40, and as Fa became smaller, D was also made smaller.

Next, a second embodiment of the present invention will be described. Since the configuration is the same as that of the first embodiment, the description thereof will be omitted, and the operation of the CPU 100 will be explained along the flowchart of FIG. 18. When a key switch (not shown) is turned on, the power is turned on and the operation starts. In step 1 (S'l), all memories are cleared to "0", and then in step νf2 (S'2), the initial value of the air-fuel ratio correction value FdO is set to 2048 (center value of 12 pits), and the torque fluctuation is Set the flag C used for size discrimination to 1. In step 3 (S'3), the master mask is set so that the CPU accepts interrupt operations, and then Stellar f4 (S'3) is set.
'4), it enters the standby state for interrupt calculation.It is always in the state of step 4 except when the interrupt calculation is deep.After that, as time passes, the timing pulse generation circuit 50 outputs
il) When the /f pulse rises from "0" to "1", the interrupt operation starts. When the interrupt operation starts, Stella 7"
10 (8'IO) disables subsequent interrupts. In step 1t (s'tl), generate Fig. 6 (6)/Gurus,
Using this signal as a trigger, the A-D converter F01 starts A-D conversion and at the same time outputs the EoC terminal as shown in Fig. 6 (I
), 4 rus becomes "1", CPU 100 becomes BU8
When the Y input becomes "1", the operation stops. This is step 12 (S'12).6 A-D converter 701
- When the D conversion is completed, the mc terminal output is shown in Figure 6 (I).
becomes "0", the CPU 100 resumes calculation. When the calculation is restarted, A-
The torque fluctuation value Tn output from the D converter 701 is read into the CPU 100. As mentioned above, the torque fluctuation value Tn is A-D converted and read at Stella f11°12.13 and counted by the intake air amount counting circuit 90 at 0 step 7'14 (8'14). In step 15 (8'15), a value 1/NE which is inversely proportional to the engine rotation speed counted by the rotation speed detection circuit 80 is read.
By taking the reciprocal of the value of m, the engine speed NE
can be found.

FIG. 14 is a map showing the peak value TA of the torque fluctuation value when the engine is operated at the best fuel economy point under each condition with NE and middle tension as parameters of the engine operating conditions. This is stored in advance in a read-only memory (hereinafter referred to as ROM). In step 16 (S'l 6 ), the read TA is searched for where the corner is located on the map shown in FIG. 14, and the TA stored in the corresponding ROM address is read out. Step 17 (S' In l7), the magnitude of the TA and Tn read in step 13 is determined, and it is determined whether the air-fuel ratio under the current engine operating condition is rich or lean compared to the best fuel efficiency point.In other words, Tn)TA If so, it is determined that the torque fluctuation is large and the air-fuel ratio is lean, and in steps 18 to 24 (8'18 to 24) the air-fuel ratio is calculated to be richly corrected according to the difference between the output value of the torque fluctuation and the target value, On the other hand, if Tn (TA), the torque fluctuation is small and the air-fuel ratio is determined to be rich, and the rich correction calculation is performed by setting the flag C for steady rich/lean discrimination to 1.In step 19, the chi correction value Pg is determined. With the calculation, the air-fuel ratio is corrected in proportion to the difference between the current torque fluctuation value Tn and the target torque fluctuation 1lTA.In other words, if the difference in torque fluctuation values is large, the air-fuel ratio is corrected to a large extent, and if the difference is small, the air-fuel ratio is corrected to a small value. The fuel ratio is corrected. In steps 20 and 21, the maximum value of the rich correction value Fa of the air-fuel ratio is set to FM. The maximum value y is determined when the engine Constant #P!, determined as an air-fuel ratio correction value sufficient to achieve the air-fuel ratio of IA Castle.Step 2
Step 2 calculates the current air-fuel ratio correction value Fd by adding the correction value Fa obtained in steps 19 to 21 to the air-fuel ratio correction value Fd obtained in the previous calculation. Steps 23 and 24 set the air-fuel ratio correction value Fd to the maximum value of 12 bits, 4095.
The maximum value is regulated. Step 25 checks steps 7 and 7 above to check whether the interrupt calculation performed last time was rich correction or lean correction. The previous calculation was 9. If it is a lean correction (when C=1), the number of divisions of the lean correction is determined in steps 26 to 28. The Stella f26 sets the flag C to 0 and counts down the lean correction counter ef if the next interrupt calculation is to continue with the lean correction calculation. Step 27 is
From the previous (b) calculation, the correction value Fa is stored in the ROM in advance.
Table 15 of FIG. L
Calculate the value of how many times to divide lean correction until
Set the counter to front b pi D with Stella f2B.

Stella 7"'29 counts down the lean correction counter t-1 when performing lean correction calculation following the previous lean correction calculation. Stella 7"'29 counts down the lean correction counter t-1. as step 32 (S
'32) The calculation result of the lean correction value FL is made not to become O. Step 32 is the calculation of the lean correction value FL, and calculates the difference in air-fuel ratio up to the target value. The correction value is divided into D arithmetic progressions obtained in step 27, and the initial correction value is made larger and later made smaller as it approaches the target value. Stay 7” 33 (8’ 33
) is the lean correction value FL from the previous air-fuel ratio correction value Fd.
The current air-fuel ratio correction value Fd is calculated by subtracting . Ste.

! 34.35 (Da 34, 35) is 5Fd(2) Minimum value set to 0 so that the air-fuel ratio correction value Fd does not become a negative number by calculation.
It is said that Step 36 (da 36) outputs the air-fuel ratio correction value Fd'1k obtained by the above calculation to the DA conversion circuit 110. Enable interrupts with steps xf and f37 (&'37), and return to the program execution state before the interrupt occurred with step and 3s (s'3g).

With the above configuration and operation, the air-fuel ratio supplied to the engine can be controlled to the best fuel efficiency point based on the detection signal from the torque detector 12. Time chart of air-fuel ratio correction calculation in the flowchart of Fig. 18 (A) Fig. 19 (A)
is the torque fluctuation signal after the filter circuit 3o, and the 19th
Figure 19 (g) shows the state of flag C for determining FiIJ, H, lean, and Figure 19 (C) shows the air-fuel ratio correction value F.
It shows an analog voltage obtained by D-A conversion of d.

If the torque fluctuation signal shown in Figure 19 is the torque fluctuation target value TA at the best fuel efficiency point, then rich correction is performed according to steps 18 to 24 of the flowchart in Figure 18, and if it is small, steps 25 to 35 Lean correction is made according to the following. Furthermore, the lean correction value Fa is proportional to the difference in torque fluctuation values, and the lean correction value FL is also proportional to the difference in torque fluctuation values.Moreover, the initial correction value is large and is reduced as it approaches the target value, so that it can be quickly Moreover, it is possible to accurately control the air-fuel ratio to achieve the best fuel efficiency.

In this embodiment, the method of determining the constant for the chi correction value Fa is the same as in the first embodiment. Regarding the lean correction value FL, find the appropriate number of lean correction divisions from the previous rich correction value Fa from the table in Figure 15, and calculate the air-fuel ratio proportional to the difference between the torque fluctuation value and the target value according to the division number. The values were divided into an arithmetic progression and the lean correction values FL were determined in descending order of values. The division number is set so that the target value is reached in 20 seconds even when the air-fuel ratio changes by 0.5, which is the maximum air-fuel ratio change.
40, and as Fi becomes smaller, D also becomes smaller.

As described above, in the present invention, torque fluctuations caused by combustion fluctuations in the engine are detected, and the detection signal is inputted to the control circuit through a filter circuit having filter characteristics in the frequency band. By extracting only a large detection signal caused by combustion fluctuations, combustion fluctuations can be accurately grasped, and the detection signal is also larger than the torque fluctuation control target value calculated based on the detection signal of the engine sensor. If the detection signal is smaller than the control target value, the supplied air-fuel ratio is corrected rich in proportion to the difference between the detection signal and the control target value, and if the detection signal is smaller than the control target value, the lean correction is performed in multiple steps (one time By keeping the correction amount the same or changing it according to the difference from the target value, the torque fluctuation value of the engine is always set to the control target value in the engine operating state. In addition, when the sensor output value is smaller than the target value (when the A/F is slightly lower than the misfire limit value), the error correction is gradually performed multiple times. Then, if the correction amount is divided into multiple times, the first correction amount is the largest, and the correction amount decreases as it approaches n times. According to the present invention, which performs correction to quickly and more accurately control to the target value, engine combustion fluctuations are determined from torque fluctuations, and the supply to the engine is performed so that engine combustion fluctuations become a certain constant value. By controlling the air-fuel ratio, it is possible to always maintain the air-fuel ratio at the best fuel efficiency point for the misfiring castle, thereby achieving a significant improvement in fuel efficiency.

[Brief explanation of the drawing]

Fig. 1 is a characteristic diagram showing the relationship between air-fuel ratio, torque fluctuation, fuel consumption rate, and exhaust gas components, Fig. 2 is a diagram showing the configuration of an embodiment of the present invention, and Fig. 3 is a control unit shown in Fig. 2. professional,
Figure 4 is an electric circuit diagram showing the timing/failure generation circuit of the control unit y) in Figure 3, Figure 5 is an electric circuit diagram showing the peak hold circuit of the control unit y) in Figure 3, 6th
The figure is a waveform diagram of each part to explain the operation of the control unit in Figure 3, the fs7 diagram is an electric circuit diagram showing the A-D conversion circuit of the control unit in Figure 3, and Figure 8 is the rotation of the control unit in Figure 3. An electric circuit diagram showing the speed detection circuit, FIG. 9 is a waveform diagram of each part to explain the operation of the rotation speed detection circuit shown in FIG. 8, and #! 1
The 0 figure is! Fig. E3 is an electric circuit diagram showing the intake air amount counting circuit of the control unit, Fig. 11 is a waveform diagram of each part to explain the operation of the intake air amount counting circuit shown in Fig. 1θ, and Fig. 12 is! The electric circuit diagram showing the D-A conversion circuit of the control unit y in Fig. 3, and Fig. 13 is the 1'j! Microcombi in Example - Massi's diagram showing the target value of torque fluctuation under each engine condition used for the calculation of the microcomputer, shown in Figure 15 and Figure 13, is shown in the evening performance. FIG. 16 is a time chart showing the air-fuel ratio correction to the engine in the first embodiment;
FIG. 17 is a characteristic diagram showing the relationship between the engine torque fluctuation value and the peak hold circuit output voltage in this embodiment, and FIG. 18 is a calculation flowchart showing the calculation order of the microcomputer in the second embodiment of the present invention. FIG. 19 is a time chart showing air-fuel ratio correction to the engine in the second embodiment. Chamber, 7--Throttle valve, 8--Electromagnetic fuel injection valve, 9--Exhaust valve, DESCRIPTION OF SYMBOLS 10... Exhaust pipe, 11... Fuel control unit, 12... Ignition coil, 13... Torque detector, 20... Amplifier, ao... Pantone filter, 4°... Black circuit, 50... Ting pulse generation circuit, 60... Peak hold circuit, 70... A-D conversion circuit, 80... Rotation speed detection circuit, 90... Intake air amount counting Circuit, 100... Correction calculation circuit, 110... D-A conversion circuit, 120...
・Patent applicant for fuel quantity calculation circuit Japan Auto Parts Research Institute Co., Ltd. Meyu Jidosha Kogyo Co., Ltd. Patent application agent Akira Aoki Patent attorney Kazuyuki Hyakutate Patent attorney Akira Yamaguchi Figure 1 kgom outside Fig. 10 90 Fig. 11 tOh t2 340- Fig. 12 ++n Fig. 14 Fig. 16 Fig. 17 0+□tw ri change W+k ""kg-" Fig. 19 Reno Fa- 1・
(Tn-TA) Continued from page 1 0 Inventor: Takashi Shigematsu, 1 Toyota-cho, Toyota-shi Toyota Motor Corporation Applicant: Toyota Motor Corporation, 1-Toyota-cho, Toyota-shi

Claims (1)

[Claims] 1. An internal combustion engine comprising: an engine senna that detects the operating state of the internal combustion engine; and a control circuit that controls the air-fuel ratio supplied to the internal combustion engine based on a detection signal from the engine senna. The fuel ratio control device includes a torque fluctuation detector that detects combustion fluctuations in the internal combustion engine as torque fluctuations, a filter circuit connected to the torque fluctuation detector and having filter characteristics in a frequency band, and an output from the filter circuit. A control target value for the combustion fluctuation of the internal combustion engine is calculated based on the detection signal of the engine sensor as an input, and this target value and the output value from the filter circuit are compared, and the target value and the output value are calculated. An air-fuel ratio control device for an internal combustion engine, characterized in that the control circuit is configured to correct the air-fuel ratio supplied to the internal combustion engine according to the difference between the air-fuel ratio and the air-fuel ratio. 2. The air-fuel ratio control device for an internal combustion engine according to claim 1, wherein the control circuit includes means for detecting a peak value of the output value from the filter circuit within a predetermined period of time. 3. If the output value from the filter circuit is larger than the target value, the air-fuel ratio supplied to the internal combustion engine is richly corrected, and the correction amount for this one time is calculated between the output value and the target value. and when the output value is smaller than the target value, the air-fuel ratio supplied to the internal combustion engine is corrected to lean, and the correction amount for that one time is corrected at least twice or more. The air-fuel ratio control device for an internal combustion engine according to claim 1 or 2, wherein the control circuit is configured to be divided so as to reach an air-fuel ratio that obtains the target value. 4. If the output value is smaller than the target value, divide it into at least two corrections, and once mb
4. The air-fuel ratio control device for an internal combustion engine according to claim 3, wherein the control circuit is configured so that the correction amounts are the same. 5. When the output value is smaller than the target value, the correction is divided into at least two times, and the amount of correction for one correction is determined according to the difference between the output value and the target value. The air-fuel ratio control device for an internal combustion engine according to claim 3, wherein the control circuit is configured to change the air-fuel ratio.