JPH0412172A - Control device for ignition timing of engine - Google Patents

Abstract

PURPOSE:To surely prevent generation of knock in the operating territory in which knock feedback control is stopped by correcting an ignition timing based on a learning value of ignition timing in a feedback control territory, and correcting it on the delay angle side at incomplete learning in the control territory. CONSTITUTION:In the device in which an ignition timing instituted based on a parameter of operating condition is corrected by feedback of detection of knock generation and a learning value, a learning completion judging means M1 is provided for judging whether learning of ignition timing in a knock feedback control territory (a) is completed or not. When knock feedback control of ignition timing is stopped by a knock feedback control stopping means M2 in a prescribed operating territory (b), ignition timing outside the knock feedback control territory is corrected based on the ignition timing learning value in the knock feedback control territory at judging completion of learning. Meanwhile, an ignition timing correcting means M3 by which at judging incompletion of learning an ignition timing outside the knock feedback control territory is corrected on the delay angle side of that at completion of learning.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an ignition timing control device for an engine that prevents knock from occurring in an operating range where knock detection is difficult.

[Prior Art] Recently, a technology that detects initial knock caused by abnormal engine combustion and optimally controls ignition timing has been widely adopted.

In this case, engine knock is often caused by pressure vibrations caused by abnormal combustion of the air-fuel mixture being detected as mechanical vibrations by an acceleration sensor attached to the cylinder head or cylinder block. It becomes difficult to distinguish it from its own vibration component.

Therefore, knock detection is stopped in the engine high speed range,
The ignition timing is corrected based on the learned value learned in the operating range where knock detection is possible. A technique is disclosed in which a value is stored and when knock feedback control is stopped in a predetermined operating range, the basic ignition timing is corrected using the stored correction value.

[Problem to be solved by the invention] However, the occurrence of knocking is closely correlated with the octane number of the fuel. For example, when fuel with a low octane number is refilled during refueling, new learning may occur while the octane number of the entire fuel decreases. If you drive at high speed before the engine is executed, the ignition timing will be corrected based on the previous learning results because the learning results up to that point will be stored and the ignition timing will be over-advanced and knocking will occur frequently. , may cause damage to the engine.

To deal with this, for example, a technology that clears the learning value every time the ignition switch is turned off and performs learning from a blank slate each time, or a sensor that detects the octane number of the fuel or a sensor that detects the refueling of the fuel can be used. Techniques that add methods are conceivable, but in the former, the learning effect is halved and improvements in engine output performance and fuel efficiency cannot be expected, while in the latter, there are problems such as increased costs. .

[Object of the Invention] The present invention has been made in view of the above circumstances, and it is an object of the present invention to prevent the occurrence of knock due to a change in the octane number of the fuel without adding any special detection means, even in an operating range where knock detection is difficult. The purpose of the present invention is to provide an engine ignition timing control device that can perform the following functions.

[Means for Solving the Problems] In order to achieve the above object, the engine ignition timing control device according to the present invention sets the ignition timing based on operating state parameters, and performs feedback correction on the ignition timing by detecting occurrence of knock. In an engine ignition timing control system that corrects the timing using learned values, as shown in Fig. 1,
learning completion determination means M1 for determining whether learning of ignition timing in the knock feedback control region is completed;
knock feedback control stopping means M2 for stopping knock feedback control of ignition timing in a predetermined operating region;
When the knock feedback control of the ignition timing is stopped by the knock feedback control stop means M2, when the determination result of the learning completion determination means M1 is that learning is complete, the knock feedback control is performed based on the ignition timing learning value in the knock feedback control area. Knock feedback that corrects the ignition timing outside the area and, when the determination result by the learning completion determining means M1 is that learning is not completed, corrects the ignition timing outside the knock feedback control area to be retarded than when learning is completed. The control area ignition timing correction means M3 is also provided with an ignition timing correction means M3 outside the control area.

Further, the learning completion determining means M1 determines that learning is complete when the amount of change in the ignition timing learning value in the knock feedback control region is within a preset range, and the learning completion determining means M1 determines that learning is complete when the amount of change in the ignition timing learning value in the knock feedback control region is changed. When the amount is not within a preset range, it is determined that learning is not completed, or when the frequency of knock occurrence in the knock feedback control area is below a preset frequency, it is determined that learning is complete, and the knock feedback control area is determined to be complete. When the knock occurrence frequency exceeds a preset frequency, it is determined that learning has not been completed.

[Function 1] In the engine ignition timing control device having the above configuration, the learning completion determining means M1 determines whether learning of the ignition timing in the knock feedback control region is completed;
When the knock feedback control of the ignition timing is stopped in a predetermined operating region by the knock feedback control stop means M2, the ignition timing outside the knock feedback control region is changed to knock feedback according to the determination result of learning completion by the learning completion determination means M1. It is corrected by the out-of-control area ignition timing correction means M3.

That is, in the operating region where the knock feedback control is stopped, when learning has been completed in the knock feedback control region, the ignition timing is corrected based on the ignition timing learned value in the knock feeder and knock control regions, and When learning is not completed in the knock feedback control region, the angle is corrected to be later than when learning is completed.

Further, the learning completion determination means M1 determines whether learning is completed based on the amount of change in the ignition timing learned value in the knock feedback control region, and determines whether the amount of change in the ignition timing learned value is within a preset range. It is determined that learning has been completed when the learning is completed, and learning is not completed when the amount of change in the ignition timing learning value is not within a preset range.

Alternatively, in the learning completion determination means Ml, it is determined whether learning is completed based on the knock occurrence frequency in the knock feedback control area, and when the knock occurrence frequency is equal to or less than a preset frequency, learning is completed, and the knock occurrence frequency is When the frequency exceeds a preset value, it is determined that learning has not been completed.

[Embodiments of the invention] Hereinafter, the present invention will be described in detail with reference to the drawings.

Figures 2 to 6 show the first embodiment of the present invention, Figure 2 is a schematic diagram of the engine control system, Figure 3 is a flowchart showing the ignition timing control procedure, and Figure 4 is a learned value. Flowchart I~ showing the update procedure, FIG. 5 is a flowchart showing the knock feedback correction value calculation procedure, and FIG. 6 is a flowchart showing the learning completion determination procedure.

(Configuration of engine control system) In Fig. 2, reference numeral 1 is the engine body;
An intake pipe 3 and an exhaust pipe 4 are connected to the intake bow J-2a and the exhaust bows 1 to 2b formed in the cylinder head 2 of the engine main body 1, and the cylinder head 2 has its ignition part connected to the combustion chamber. A spark plug 5 exposed at 1a is attached.

An air cleaner 6 is connected to the upstream side of the intake pipe 3,
Immediately downstream of this air cleaner 6, an intake air 1 sensor (in the figure, a wire air flow meter) 7 is installed.

Further, a throttle valve 8 is interposed in the middle of the intake pipe 3, and a throttle opening sensor 9 is connected to the throttle valve 8. Furthermore, downstream of the throttle valve 8 of the intake pipe 3, A pressure sensor 10 is facing.

Further, an injector 11 faces directly upstream of the intake boat 2a communicating with the combustion chamber 1a, and an 02 sensor 13 faces directly upstream of the catalytic converter 12 interposed midway in the exhaust pipe 4.

In addition, the cylinder block 1b of the engine body 1 is
A cooling water temperature sensor 14 faces the cooling water passage 1c formed in the cylinder block 1b, and a knock sensor 15 is attached to the side wall of the cylinder block 1b.

Further, a crank rotor 16 is fixed to the crankshaft 1d of the engine main body 1, and this crank rotor 16
A crank angle sensor 17 is provided on the outer periphery of the crankshaft.

(Circuit configuration of control means) On the other hand, reference numeral 20 is a control device (ECTJ) consisting of a microcomputer, etc., and the CPU of this ECU 2.0
(Central processing unit 21, ROM22, RA, M2
The S, backup RAM 23a, and I10 interface 24 are connected to each other via a pass line 25, and set the ignition timing based on the operating state parameters, detect the occurrence of knock, perform feedback compensation on the ignition timing, and adjust the ignition timing based on the learned value. The function of correcting ignition timing control is realized, and other functions such as fuel injection control are also realized.

Each of the sensors 7.9, 10.13.1415.17 is connected to the input port of the I10 interface 24, and the injector 11 is connected to the output port of the I10 interface 24 via the drive circuit 26. At the same time, the igniter 27 is also connected.

An ignition coil 28 is connected to the igniter 27, and the ignition plug 5 is connected to the ignition coil 28 via a distributor 29.

The ROM 22 contains a control program, a basic ignition timing map MPBSE, and an optimum ignition timing map M, which will be described later.
Fixed data such as P HBT is stored, and the learning map M P IGL is stored in the backup RAM 23a.
RN is formed and the ignition switch (not shown)
Data is retained even after it is turned off. Further, the RAM 23 stores data from each of the sensors, data processed by the CPU 21, and various flags.

In the ECU 20, when an ignition switch (not shown) is turned on, the CPU 21 immediately executes initialization and clears various flags. When the engine starts operating, the CPU 21 calculates fuel injection tTi, ignition timing θIG, etc. based on various data stored in the RAM 23 in accordance with the control program stored in the ROM 22.

Then, the fuel injection time and ignition time are set in counters and timers for control timing setting, and the drive pulse width signal for the injector 11 and the ignition signal for the igniter 27 corresponding to the calculated fuel injection amount Ti are set at the optimal timing. and output it.

(Operation) Next, the operation of the embodiment with the above configuration will be explained based on flowchart 1 of FIGS. 3 to 6.

(Ignition timing control procedure) Flowchart 1 shown in FIG. In step 5102, the intake pipe pressure PM is calculated from the output signal of the pressure sensor 10, and the process proceeds to step 5103.

In step 5103, the cooling water temperature T- is calculated from the output signal of the cooling water temperature sensor 14, and in step 5104, the presence or absence of knocking is determined from the output signal of the knock sensor 15.
If there is knocking, a knocking determination flag FLAG1 is set to identify the occurrence of knocking.

Next, the engine speed N calculated in step 5101 above and the intake pipe pressure PM calculated in step 5102 above.
In step 5105, the optimum ignition timing map M P HBT is searched using the parameter , and the optimum ignition timing is determined to give the maximum allowable advance difference in the knock limit of the octane number of the fuel set to be used in the engine, for example, premium gasoline. While calculating MBT, the basic ignition timing map MPBSE is searched in step S+06 to calculate the basic ignition timing IGBSE in the current operating state.

Next, step 5i07.3108. At S+09, it is checked whether the current operating state satisfies the conditions for performing learning. That is, in step 5107, the engine rotation speed N reaches the learning lower limit rotation speed N1. OW (for example,
500 rpm) and learning upper limit rotation speed N tlPPER (
For example, the intake pipe pressure PM and the set value P5 are checked in step 8108.
ET (e.g. 900 if the engine is supercharged)
mmh) to check whether the condition of PH>PSE is satisfied.

Also, in step 5109, the cooling water temperature T- and the set value TW
SET (e.g. 70°C) and TWS
It is checked whether the ET conditions are satisfied or not, and if all the conditions in step 5107.5i08.5109 are satisfied, the subroutine for updating the learning value is called in step 3110, and the learning value is updated. The process proceeds to step 5111, while the above step 3107.51
If there is even one step in which the conditions are not satisfied in 08 and 3109, step 5110 is jumped and the process proceeds to step 5111 without updating the learning value.

In step 5111, the engine speed N and the learning upper limit speed NUPPER are compared, and N≦N IIPPE is determined.
If R(7), the process proceeds to step 5112, where the overall learning value A, TOT, is used to globally correct the ignition timing in all regions.
, the region-specific learning value A PRT for correcting the ignition timing in a partial region, and the value obtained by adding the region-specific learning value A PRT to the step 5105 described above.
Compare with the optimum ignition timing MBT calculated in .

Then, in step 5112 above, MBT≦ATOT+A
In the case of PRT, the process proceeds from step 5112 to step 5113, where the optimum ignition timing MBT calculated in step 5105 is set as the learning ignition correction value IGLRN (I GLRN-MBT>actual ignition timing is less than the optimum ignition timing MBT). Avoid advancing.

On the other hand, in step 5112 above, M BT>A, TOT
In the case of +-APRT, the process proceeds from step 5112 to step 5114, where the sum of the overall learning value A, TOT, and the area-specific learning value APRT is determined as the learning ignition correction value IG.
As LRN I GLRN -ATOT +APRT
), proceed to step 5122.

When the process proceeds from step 5113 or step 5114 to step 5122, a knock feedback correction value calculation subroutine is called to obtain the knock feedback correction value AKN, and in step 5123, the basic ignition timing IGBSE calculated in step 5106 is
The above step 5113 or the above step 5114
The learned ignition correction value IGI-RN set in step 5122 is added, and the knock feedback correction values A and KNK obtained in step 5122 are added to determine the final ignition timing I.
Calculate GT IGT←IGBSE+IGLRN+AK
N), exit lretin.

On the other hand, in step 5111 above, N > N UPPER
In this case, steps 5111 to 511 above are performed.
Proceed to Step 5, find the minimum value of the learning value A PRT for each area, that is, the minimum learning value APRTmin for all areas, and proceed to Step 5.
At step 116, the value of a learning completion flag FLAG3 set by a subroutine for determining completion of learning, which will be described later, is checked to determine whether or not learning has been completed.

If it is determined in step 8116 that learning has been completed, the process proceeds to step 5117, where the overall learning values A, T
The minimum learning value APR obtained in step 5115 above is applied to OT.
After adding Tmin, subtract the first set value AR to set the learned value A LRN at high engine speeds that are outside the knock feedback control area. A LRN←A T
OT +A PRTmin -A RTl >, if the learning is not completed, proceed to step 8118, add the minimum learning value A PRTmO obtained in step 5115 to the overall learning value ATOT, and then add the minimum learning value A PRTmO obtained in step 5115 above to the above first set value. A
A second setting value ART2 (
Subtract A, RTl <ART2> and set the learning value A LRN at high engine speed (ALRN ←AT
OT +APRT sin -ART2).

That is, when setting the learning value A LRN at high engine speeds outside the knock feedback control area, the ignition timing correction amount is switched depending on whether learning in the knock feedback control area is completed, and if learning is not completed, the learning value is changed. The amount of correction is made smaller than that at the time of completion, and the amount of advance is made smaller, thereby preventing the occurrence of knocking.

In this case, the minimum learning value AP is added to the total learning value A TOT.
The value obtained by adding RTmin may be multiplied by 1 for the first I series coefficient and 2 for the second coefficient (K2<Kl<1), switching between them depending on the learning completion determination result.

Then, in step 5119, the learning value ALR set in step 5117 or step 8118 is determined.
Compare N with the optimum ignition timing MBT and find that At-RN≧MB
In the case of T, in step 5120, the learned ignition correction value IGI-
Set the optimum ignition timing MBT to RN (I G LR
N←MBT) Proceeds to step 5122, while ALRN
In the case of <MBT, in step 5121, set the learned value A LRN to the learned ignition correction value IGLRN.
GLRN←ALRN), the process similarly proceeds to step 5122.

Then, similarly, the subroutine for calculating the knock feedback correction value is called, and in step 5123, the final ignition timing TGT is calculated, and the program exits.

(Learned Value Update Procedure) Next, a learned value update program will be described based on flowcharts 1 to 1 in FIG. 4.

In the above ignition timing control program, step 5
When all the conditions of 107.5108.3109 are satisfied and the learning condition is established, a learning value update subroutine is executed.

In this learning value update subroutine, the backup RA
Learning map of M23a MP By implementing this, the frequency of knock occurrences is reduced at an early stage.

That is, it is determined in step 5201 whether or not the entire learning has been completed, and if the entire learning has not been completed, the address of the entire learning value ATOT is saved in the register X in step 5202, and the entire learning is completed. If the process has shifted to area-specific learning, the process advances to step 5203 and the address of the area-specific learning value APRT is saved in register X.

Next, in step 5204, the knock occurrence determination flag FL
AG1 checks whether there is knocking, and if there is no knocking,
Proceed to step 5205 and calculate the value obtained by adding the area-specific learning value A PRT to the overall learning value A TOT and the optimum ignition timing MBT.
Compare with.

In step 5205 above, MBT≦A TOT +A
At the time of PRT, step 521 is performed without updating the learning value.
3, and when MBT>ATOT+APRT, the process proceeds from step 5205 to step 8206, where timer T1 is read from timer 2 and the duration of the no-knock state is checked.

Then, when the timer T1 in the timer TH1 is less than or equal to a certain time TSET1 (for example, 1 sec >
Similarly, step 5213 jumps without updating the learned value, and the timer T1 clocks T1 for a certain period of time TSET.
1, that is, when no knock occurs for a certain period of time, steps 8206 to 32 are performed.
Proceeding to step 07, the value of the knock feedback correction value AKNK is checked.

In the above step 5207, when AKN≠O, similarly jump to step 5213, and when A,KNK=O, in step 5208, the contents of the memory specified by register X, that is, the contents saved in step 5202 or step 5203 The learning value (X) stored in the memory of the address is given a certain advance angle amount A ADV
is added and updated to the advance side ((X) ← (X) 10AADV), and in step 5209, the timer TH1 that measures the no-knock state is cleared, and the process proceeds to step 5213.

As a result, when the learning conditions are satisfied and there is no knocking, when all the conditions in steps 205 to 5207 are YES, a constant advance amount A ADV is generated at regular intervals.
is added and the learned value is updated in the advance direction.

On the other hand, when there is a knock in the upper step 5204,
Proceeding from step 5204 above to step 5210, subtract a certain retard amount A RET from the memory content (X) specified by register X (X) - (X,)AR
ET), the learning value is updated to a constant star retard side for each knock.

Next, in step 5211, a timer TH1 and a timer TH2, which will be described later, are cleared, and in step 212,
The counter CTI is incremented to count the number of knock occurrences, and the process advances to step 5213.

Next, when the process advances to step 5213 and subsequent steps, the conditions for ending the overall learning are checked, and first, in step 5213, it is checked whether or not the overall learning is in progress. Then, if the overall learning is completed, return as is and move on to area-specific learning; if the overall learning is not completed, step 5 above.
The process proceeds from 213 to 5214 to determine whether the overall learning value A TOT has reached a predetermined constant A HAX.

In step 5214, the overall learning value A TOT does not reach the predetermined maximum advance angle A WAX.
In the case of HAX, proceed to step 5215 and obtain the overall learning value A.
Clear timer 042 to measure the cumulative time when TOT is greater than or equal to the maximum advance angle AHAX, and step 5
At step 217, the count value C1 of the counter CT1, that is, the number of knock occurrences is read out, and this number of knock occurrences C1 is compared with a predetermined set number of times C3ET1 (for example, 5 times).

In step 5217, if CI < C3ET1, that is, the number of knock occurrences 01 has not reached the set number of times C5ETI, the routine is exited and returned, and C1≧C
5ET1, that is, when knocking occurs a set number of times, for example, 5 times, it is determined that the entire learning is completed. Then, in step 8218, the overall learning end flag FLAG2 is set to 1~, exits the routine, returns, and shifts to area-specific learning.

On the other hand, in step 5214, the overall learning value A TOT reaches the maximum advance angle A HAX, and ATOT ≧ A HAX.
If no, steps 5214 to 82 above are performed.
Proceed to step 16 and the overall learning value A TOT is the maximum advance angle A above.
Timer TH that measures cumulative time in a state equal to or higher than HAX
Read the value of 2 and enter the cumulative time T2 and set time TSET2 (
For example, when T2<TSET2, the process proceeds to step 5217 to check the number of knock occurrences from the count value of the counter CT, and when T2≧TSET, the process proceeds to step 8218 to set the overall learning end flag F. 1. Set AF 2, return, and move on to area-based learning.

(Knock feedback correction value calculation procedure) FIG. 5 is a flowchart showing the knock feedback correction value calculation procedure, step 51 of the above-mentioned ignition timing control program.
When this subroutine is called in step 22, first, in step 5301, it is determined whether the engine rotation speed N is higher than the learning upper limit rotation speed NUPPER.

In step 5301 above, when N>N IJPPER, the process proceeds to step 5302 and clears the knock feedback correction value AKNK to O (A KNK←
0), the knock feedback control is stopped and the process returns, and when N≦N UPPER, step 5 is executed.
Proceed to step 303 to check whether knocking has occurred.

If no knock occurs in step 5303, the process proceeds to step 5304, where timer TH3 measures the duration of the no-knock state in the same way as timer TH1 described above.
When T3≦T5ETI, exit the routine and return. When T3>TSE1, that is, when no knock occurs for a certain period of time, proceed from step 5304 to step 5305 to perform knock feedback correction. AKNK (-A
, KNK +AADV1), the timer TH3 is cleared in step 5307, and the process returns.

On the other hand, if it is determined in step 5303 that knock has occurred, the process proceeds from step 5303 to step 8306, where the knock feedback retard amount A RETl is subtracted from the knock feedback correction value AKNK and updated.
KNK - AKNK - ARETl) - Step 530
At 7, the timer Tl43 is cleared and the process returns.

The knock feedback correction value AKN at this time is 0.
is counted as an upper limit, and when a knock occurs, the learned ignition correction value IGLRN and the knock feedback correction value AKNK are simultaneously updated in the retard direction by the learned value update routine and the knock feedback correction value calculation routine.

Then, when knocking stops occurring, the feedback correction value AKNK is first updated in the advance direction, and when the feedback correction value AKNK is advanced to 0, it does not advance any further, and then the learned ignition correction value ■GLRN is updated in the advance direction.

(Learning Completion Judgment Procedure) On the other hand, for the cumulative time during which the engine operating state is within the learning completion determination area, learning completion is determined every time this cumulative time or a set time is reached, and learning is determined to have been completed. and,
A learning completion flag FLAG3 is set.

This learning completion flag FLAG3 is stored at a predetermined address in the RAM 23 and held while the engine is running, but is initialized and cleared when the engine is started.

Hereinafter, the procedure for determining completion of learning will be explained based on the flowchart shown in FIG.

First, in steps 3401.5402, the intake pipe pressure P)l
and engine rotation speed N, it is checked whether the current operating state is within the learning completion determination area, and when the condition of PH>PSET is satisfied in step 5401, step 5402
If the condition is not met, exit the program and return.

Further, in step 5402, the engine rotation speed N is the lower limit rotation speed N1 (for example, 200
Orpm) and upper limit rotation speed N2 (for example, 4500r
If it is within the set range, the process advances to step 5403, and if it is outside the set range, the program exits and returns in the same way.

When it is determined that the current driving state is within the learning completion determination area and the process proceeds to step 5403, a timer TH4 that measures the cumulative time within this determination area is counted up, and in step 5404, the timer TH4 is counted up. The clock T4 is read out and the set time TSET4 (for example, 10 seconds) is read out.
c).

If T4≦T5ET4 in step 5404, the program exits the program, and if T4>TSET4, that is, if the cumulative time within the determination area exceeds the set time, the program exits from step 5404 to step 54.
Proceed to 05, and the amount of change ΔL in the learned value during that period is within the setting range (Ll<ΔL<L2; For example, -2°
degree range).

In step 5405, when the learned value change ΔL is not within the set range, the process jumps to step 5407, while when the learned value change 1ΔL is within the set range, the process proceeds from step 5405 to step 5406, where the learning completion flag FLAG3 and step 5407
Clears timer TH4, ends the program, and returns.

Note that the learning completion determination routine may be executed every time the learning value is updated a predetermined number of times.

In other words, if it is determined that learning is complete as a result of this learning completion determination, the learning result in the knock feedback control area reflects the octane number of the fuel, and even in the driving area where knock detection is difficult, refueling etc. Knock can be prevented from occurring due to changes in the octane number of the fuel.

Furthermore, when stopping the engine and refueling, for example, if you mistakenly refill regular gasoline instead of premium gasoline, or if poor quality gasoline is mixed in in areas with poor fuel conditions, etc. Even if you suddenly start driving at high speeds while the octane number of the fuel is low,
Since the learning completion flag Fl-AG3 is cleared, learning is not completed and the backup RAM 23 a
The ignition timing is not corrected based on the learning result when the octane value of the fuel stored in the engine is high, and the ignition timing is corrected using a preset value, thereby preventing the occurrence of knocking.

(Second example) Next, a second embodiment of the present invention will be described.

This second embodiment differs from the first embodiment described above only in the procedure for determining completion of learning, and the determination of completion of learning is performed based on the frequency of knock occurrence.

(Learning completion determination procedure) The engine ignition switch (not shown) is turned on.
Then, the completion of learning is determined based on the number of knock pulses every set time within the learning completion determination area, and the procedure will be described below based on flowcharts 1 to 7 shown in FIG.

First, in steps 5501 and 5502, it is checked from the intake pipe pressure PH and the engine speed N whether or not the current operating state is within the learning completion determination area.
When the condition H>PSET is satisfied, the process advances to step 5502, and when the condition is not satisfied, the program exits and returns.

Further, in step 5502, the engine rotation speed N is the lower limit rotation speed Nl (for example, 200
0rl) III) and upper limit rotation speed N2 (for example, 45
If it is within the set range, the process advances to step 5503, and if it is outside the set range, the program exits and returns in the same manner.

When it is determined that the current driving state is within the learning completion determination area and the process proceeds to step 5503, a timer IH4 that measures the cumulative time within this determination area is counted up, and in step 5504 it is determined whether or not knocking has occurred. It is determined.

If there is no knock in the above step 5504, the process jumps to step 5506, and if there is a knock, the process proceeds from the above step 5504 to step 5505 and the counter C is
The knock pulse is caused by T2, and the count value C is
Count up 2 and proceed to step 3506.

In step 8506, the time T4 of the timer TH4 is read, and compared with the set time TSET4 (for example, 10 seconds). If T4≦T5ET4, the program exits the program. When the cumulative time exceeds the set time, the process proceeds from step 3506 to step 3507, reads out the count value C2 of the counter CT2, and calculates the set number of times C5E.
Compare with T2 (eg, 4 times).

Then, the count value C2 of the counter CT2 is C2≦C
If C2>C3ET2, that is, the number of knock occurrences is less than the set number, it is determined that learning is complete, and the process proceeds to step 3508, where the learning completion flag FLAG3 is set to 1-L. In step 5509, the counter C2 and timer TH4 are cleared, the program is ended, and the program returns.

In this second embodiment, as in the first embodiment, learning is determined to be complete only when it is determined that the learned value reflects the octane number of the fuel, so knock occurrence outside the knock feedback control area is prevented. This can be prevented.

[Effects of the Invention] As described above, according to the present invention, the learning completion determining means determines whether learning of the ignition timing in the knock feedback control region is completed, and the knock feedback control stopping means determines whether or not learning of the ignition timing in the knock feedback control region is completed. When the knock feedback control of the ignition timing is stopped, when the determination result of the learning completion determination means is that learning is completed by the knock feedback control area outside ignition timing correction means, the knock is performed based on the ignition timing learning value in the knock feedback control area. To correct the ignition timing outside the feedback control region, and when the determination result by the learning completion determination means is that learning is not completed, to correct the ignition timing outside the knock feedback control region to the R negative side compared to when learning is completed. ,
Even in operating ranges where knock detection is difficult, knocking can be prevented due to changes in the octane number of the fuel due to refueling, etc., and it can be constructed at low cost without adding any special detection means. .

Furthermore, the learning completion determination means
In order to determine whether learning has been completed based on the change in the ignition timing learned value in the knock feedback control region, or to determine whether learning has been completed based on the frequency of knock occurrence in the above-mentioned knock feedback control region, learning is performed. When it is determined that the process has been completed, the learned value reflects the octane number of the fuel, and excellent effects such as the ability to effectively incorporate the learned result into the ignition timing outside the knock feedback control area are achieved.

[Brief explanation of the drawing]

Figure 1 is a block diagram showing the basic configuration of the present invention, Figures 2-
Fig. 6 shows the first embodiment of the present invention, Fig. 2 is a schematic diagram of the engine control system, Fig. 3 is a flowchart showing the ignition timing control procedure, and Fig. 4 is a flowchart 1 showing the learning value update procedure. -, Fig. 5 is a flowchart 1- showing a knock fee I/back correction value calculation procedure, Fig. 6 is a flowchart 1- showing a learning completion determination procedure, and Fig. 7 is a flowchart showing a second embodiment of the present invention. This is a flowchart without showing the learning completion determination procedure. Ml...Learning completion determination means M2 Knock feedback control stop means M3...Ignition timing correction means outside the knock detection area, r--Fig. 3, Fig. 7

Claims (3)

[Claims] (1) In an engine ignition timing control device that sets the ignition timing based on operating state parameters, detects the occurrence of knock, performs feedback correction on the ignition timing, and corrects it using a learned value, the ignition timing in the knock feedback control region is adjusted. learning completion determining means for determining whether learning is completed; knock feedback control stopping means for stopping knock feedback control of ignition timing in a predetermined operating region; and knock feedback control of ignition timing by the knock feedback control stopping means. When the learning completion determination means determines that learning is complete, the learning completion determination means corrects the ignition timing outside the knock feedback control area based on the ignition timing learning value in the knock feedback control area. and a knock feedback control area outside ignition timing correcting means for correcting the ignition timing outside the knock feedback control area to a later side than when learning is completed when the determination result in is that learning is not completed. Engine ignition timing control device. (2) The learning completion determining means determines that learning is complete when the amount of change in the ignition timing learning value in the knock feedback control region is within a preset range, and the learning completion determination means determines that learning is complete when the amount of change in the ignition timing learning value in the knock feedback control region changes. 2. The engine ignition timing control device according to claim 1, wherein when the amount is not within a preset range, it is determined that learning has not been completed. (3) The learning completion determination means determines that learning has been completed when the knock occurrence frequency in the knock feedback control region is less than or equal to a preset frequency, and when the knock occurrence frequency in the knock feedback control region exceeds the preset frequency. 2. The engine ignition timing control device according to claim 1, wherein the engine ignition timing control device determines that learning is incomplete when the learning is completed.