JPH02305371A - Method of ignition timing learning control - Google Patents

Abstract

PURPOSE:To set ignition timing in an optimum way when the ignition timing is determined by the summation of whole learning and partial learning, by suspending each learning in a high engine rotational region, and by compensating ignition timing according to the high rotational region compensation based on each learning value. CONSTITUTION:A control unit 30 discriminates a learning condition through a means 40 according to engine rotation, pressure of suction pipe and water temperature calculated by each means 31-33, and to a stated region discriminated by a means 37, and a learning value is updated by a means 41 according to the learning condition. Based on the number of knocking and spark angle detected by means 43, 44, a whole learning value is memorized by a means 42, and, a partial learning value being memorized by a means 45, learning compensation is calculated by a means 46 according to the learning values. When a high rotational region is judged by a means 53, the high rotational region compensation set by a means 52 is selected by a means 54.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an ignition timing learning control method for controlling a vehicle engine by learning the ignition timing in each operating state and setting it to an optimal value. This paper relates to a method for estimating the optimal correction amount in a high rotational speed range where knocking cannot be detected.

[Conventional technology]

In general, ignition timing learning control is performed in parallel with knocking feedback control by gradually advancing or retarding the ignition timing depending on whether or not knocking occurs, while learning the amount of correction, and determining the ignition timing at the knocking limit. There is. For this reason, it is important to accurately detect knocking, which is determined by installing a knock sensor on the engine body and selecting the knocking frequency from the vibration signal from this knock sensor. . However, in a high rotational speed range, various vibration noises such as intake/exhaust valve seating noise increase, making it difficult to distinguish signals near the knocking frequency.

Therefore, in a high rotation speed range, feedback control and learning control are interrupted or the control width is reduced in order to avoid feedback and learning due to erroneous detection of knocking. The ignition timing is often set on the safe side to prevent knocking even when various variations are taken into consideration, and as a result, the ignition timing is insufficiently advanced, leading to deterioration in output and fuel efficiency. For this reason, it is desirable to estimate the ignition timing correction amount as appropriately as possible in the high rotational speed region where knocking cannot be detected.

Conventionally, in the learning control of the ignition timing, estimating the ignition timing in the high rotation speed range has been carried out, for example, in
There is a prior art in Japanese Patent No. 164076. Here, it is shown that a correction value corresponding to the deviation between the actual ignition timing and the basic ignition timing is stored during feedback of the ignition timing, and the ignition timing is determined from this correction value while the feedback is stopped.

[Problem to be solved by the invention]

By the way, in the prior art described above, by using the correction value that is being fed back as is even while the feedback is stopped, the correction value can be divided, but the reliability of this correction value is a problem. That is, the correction values vary relatively greatly depending on changes in operating conditions and fuel octane number, and which one of them is used greatly affects engine output and occurrence of knocking. Therefore, unless the selection of the correction value is made clear, inappropriate correction values will cause inconveniences such as increased occurrence of knocking.

Here, as learning control for ignition timing, there is a method of performing overall learning common to all operating states, then performing partial learning for each operating state, and determining the correction amount by adding the learning values of both. . In this method, the overall learning value can be used as is even in the high rotational speed range, and if an appropriate value is selected from the partial learning, the correction amount in the high rotational speed range can be estimated sufficiently appropriately, which is preferable.

The present invention has been made in view of the above points, and its purpose is to optimally estimate the ignition timing under learning control interruption conditions in a high rotation speed range in learning control using overall learning and partial learning. The object of the present invention is to provide an ignition timing learning control method that enables the ignition timing learning control method.

[Means to solve the problem]

In order to achieve the above object, the ignition timing learning control method of the present invention executes overall learning in which the overall learning value is set while advancing or retarding the ignition timing depending on the presence or absence of knocking, and then, for each operating state, the ignition timing learning control method In this method, partial learning is executed to advance or retard the ignition timing depending on whether the ignition timing is present or not, and the ignition timing is corrected and determined based on the sum of these overall learning values and partial learning values. Now, interrupt the above overall learning and partial learning, calculate the high rotation speed range correction amount based on the above overall learning value and partial learning value, and correct the ignition timing in the high rotation speed range with the high rotation speed range correction amount. It is up to you to decide.

[For production]

With the above method, under conditions where knocking can be detected, overall learning and partial learning are performed in relation to the presence or absence of knocking, and the correction amount is determined while learning based on the learning values of both, and the ignition timing is optimally set near the knocking limit. It is determined. In the high engine speed range where knocking cannot be detected, the learning control is interrupted, and by using the values selected from the overall learning value and the partial learning value of the learning control, the high speed range correction amount suppresses knocking. The ignition timing is estimated to be advanced and highly reliable, and the ignition timing in the high rotation speed range is optimally determined by this correction amount.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

Referring to FIG. 2, an outline of an engine to which the present invention is applied will be described. The symbol I is the engine body, and the combustion chamber 2
An intake valve 4 is connected to the intake port 3 that communicates with the exhaust port 5.
An exhaust valve 6 is provided in the combustion chamber 2, and a spark plug 7 is provided in the combustion chamber 2.
is installed.

An air cleaner 8 serves as an intake system, and an intake pipe 9. A throttle body II having a throttle valve 10 communicates with the intake port 3 via an intake manifold 12, and an exhaust port 5 as an exhaust system communicates with an exhaust pipe I3. A bypass passage 15 is provided, and an injector 16 is mounted in a multi-point manner at the inlet of the suction port 3.

A crank angle sensor 20 is used as a control system. Throttle valve IO
It has a downstream pressure sensor 21, and these sensor signals are input to the control unit 30 to detect the engine rotational speed Ne from the crank angle and the intake pipe pressure Pm as an engine load from the pressure downstream of the throttle valve 10. Basic fuel injection amount T depending on engine speed Ne and intake pipe pressure Pm
Determine p. In addition, the water temperature Tw of the water temperature sensor 22, the intake air temperature Ta of the intake air temperature sensor 23, the signal of the O2 sensor 24, etc. are also input to the control unit 30, and based on these signals, various corrections are made to the basic fuel injection amount Tp to determine the fuel injection amount. Ti is calculated and a fuel injection signal number having a pulse width corresponding to the fuel injection amount TI is outputted to the injector 16 to inject fuel according to each operating state. The opening degree of the throttle valve 10 is detected by the throttle opening sensor 25, and when the throttle opening (or idle switch) determines that the engine is idling, the opening degree of the idle control valve 14 is adjusted.
The engine speed Ne is feedback-controlled to a predetermined idle speed. Furthermore, for the ignition system, the basic ignition timing, advance limit, etc. are determined based on the engine speed Ne and suction pipe negative pressure Pm mentioned above, and the retardation or advance is determined depending on whether or not knocking is detected by the knock sensor 26 attached to the engine body 1. , and learn and determine the optimal ignition timing IGT.

Then, an ignition signal corresponding to the optimum ignition timing IGT is output to the ignition plug 7 via the igniter 271, ignition coil 28° distributor 29, and ignition and combustion occur at a predetermined crank angle before the piston top dead center.

Referring to FIG. 1, the ignition timing learning control system for the above-mentioned ignition system will be described.

First, the ignition timing setting control system will be described.The crank angle sensor 20. Pressure sensor 21. The signals from the water temperature sensor 22 and the knock sensor 26 are transmitted to the engine rotation speed calculation means 31 of the control unit 30. Suction pipe pressure calculation 'Means 32.
Water temperature calculation means 33. input into the knocking determination means 34,
Engine speed Ne, suction pipe pressure Pm.

Obtain the water temperature Tw and the presence or absence of knocking. The engine speed Ne and intake pipe pressure Pm are determined by the basic ignition timing search means 35.
The basic ignition timing IGB and the basic ignition timing IGBT (basic ignition timing IGBT) are input to the advance limit search means 36 and
(set by the advance angle amount for B) is searched using the basic ignition timing map and the advance angle limit map.

These two maps are set based on, for example, the characteristics shown in FIG. 4, and the advance limit MBT becomes smaller as the engine speed Ne increases. The basic ignition timing IGB corresponding to the knocking limit of regular gasoline is set to the advanced side as the engine speed Ne increases, and the knocking limit shifts in parallel to the advanced side as the octane number increases. In addition, the advance limit MBT and the basic ignition timing IGB have similar characteristics with respect to engine loads such as intake pipe pressure Pm, and these engine speed N
Advance angle limit MBT and basic ignition timing IGB corresponding to each operating state are searched on a map using the parameters e and suction pipe pressure Pn+.

The advance limit MBT and the learned correction amount IGL (set by the advance amount with respect to the basic ignition timing IGB), which will be described later, are input to the region determining means 37, and the two are compared as shown in FIG. Area Da where the limit MBT can be obtained (MBT≦I
GL), area Db (MB T
>IGL), and the result of this determination is input to the ignition timing calculation means 38.

The advance limit MBT, basic ignition timing IGB, learning correction amount IGL, and knocking feedback correction amount AK, which will be described later, are input to the ignition timing calculation means 38, and the ignition timing IGT is calculated and determined as follows.

I GT= I GB+ I GL+AKHere, MB
In the region where the advance angle limit MBT is T≦IGL, the optimum ignition timing can be set by using the value of the advance limit MBT as is. Therefore, in the region where MBT≦IGL, MBT is set instead of IGL. Ignition timing IGT is calculated by substituting .

The value of the ignition timing IGT thus calculated and the crank angle signal of the crank angle sensor 20 are the igniter drive means 39
is input, and an ignition signal is output at a crank angle corresponding to the ignition timing IGT.

Next, the ignition timing learning value update control system will be described.
The above-mentioned rotational speed Ne, suction pipe pressure Pi, water temperature Tw
and a learning condition determining means 40 into which the region determination result of the region determining means 37 is input. Here, the current operating state is an operating state (low load side) in which knocking can be detected with high accuracy after warm-up based on the engine speed Ne, suction pipe pressure Pm, and water temperature Tw.

(excluding extremely low rotation speed side and high rotation speed side), and MBT
>I It is determined that the learning condition is satisfied in the case of the region Db where the advance angle limit MBT of GL cannot be obtained, and this determination result is input to the learning value updating means 41.

The learning value updating means 41 receives the operating status of the engine speed Ne and the suction pipe pressure Pm, and a signal indicating the presence or absence of knocking from the knocking determination means 34, and performs overall learning or partial learning depending on the determination result that the learning conditions are satisfied. Execute selectively. First, when starting the engine, overall learning is performed, and the overall learning value AT is read out from the overall learning value storage device 11 and stage 42, and the learning value AT is updated depending on the presence or absence of knocking, and the learning value AT corresponds to the octane number of the fuel. approximate the knocking limit. In other words, if there is no knocking, the overall learned value AT is updated to the advanced angle side at a fixed rate at regular intervals, and if knocking occurs, the overall learned value AT is updated to the retarded angle side at a fixed rate every time knocking occurs. Update to. Then, the number of knockings and the amount of advance of the overall learned value AT in this case are detected by the number of knocking detection means 43 and the amount of advance angle detection means 44. When the maximum advance amount AM is reached, it is determined that the knocking limit has been approximated, the overall learning value AT at this time is stored, and the overall learning is completed.

In addition to the above-mentioned overall learning, as partial learning, a partial learning value storage means 45 is provided with detailed addresses for each driving state.
The learning value updating means 41 reads the partial learning value AP from the partial learning value storage means 45 after the above-mentioned overall learning is completed.

Then, for each operating state, the ignition timing is adjusted accurately for the partial learned value AP while updating the partial learned value AP to the advance or retard side in relation to the presence or absence of knocking in the same way as described above. It is often set near the knocking limit. These overall learning value AT and partial learning value AP are input to the learning correction amount calculation means 46 to calculate the learning correction amount IGL by IGL-AT+AP. It is now used in timing control systems.

Next, regarding the knocking feedback control system, the knocking feedback correction amount setting means 4 receives the knocking presence/absence signal from the knocking determination means 34.
It has 7. This knocking feedback correction amount setting means 47 sets the retard side correction amount AK only when knocking occurs.
When knocking occurs, a retard amount γF that is several times larger than one retard amount γ of the learning control is determined, and after knocking is avoided, it is approximately the same as the advance amount a of the learning control.
The advance angle amount ap is determined, and these retardation amount γF, advance angle amount a,
Accordingly, the knocking feedback correction amount A-13=K is output to the ignition timing calculation means 38. In addition, in order to prevent mutual interference with learning control during knocking feedback control and to prevent variations in learning correction amount IGL, advance angle prohibiting means 48 is provided to which knocking feedback correction amount AK of knocking feedback correction amount setting means 47 is input. , the learning value updating means 41 is prohibited from advancing the learning correction amount IGL until the knocking feedback correction amount AK=0.

Furthermore, regarding the ignition timing estimation control system in the high rotational speed range, the partial learning value storage means 45 includes a selecting means 50, which selects any one of the partial learning values AP for high rotational speed. Here, the partial learning value AP is set as APj to APm in a map divided for each operating state according to engine speed No. and suction pipe pressure PI11 as shown in Fig. 5. When selecting for number area,
The reliability of the partial learning value on the highest rotation speed side is high. Therefore,
One method for selecting a value for a high rotational speed region from the partial learning value AP by the selection means 50 is to select the partial learning value APh (A in FIG.
The partial learning value APh and the total learning value AT are input to the correction amount calculation means 51, and the correction amount AL is calculated as AL-AT+APh-K (K is a constant). Calculated according to the following formula. The correction amount AL is further inputted to the high rotation speed range correction amount setting means 52, and compared with the advance limit MBT to find that AL<M.
In the case of BT, the correction amount AL is output as the high rotation speed range correction amount IGL.

On the other hand, it has a high engine speed range determining means 53 which inputs the engine speed Ne, and determines that the engine speed is in the high engine speed range, for example, when knocking cannot be detected when Ne > 5000 rpm.

Then, either the learning correction amount IGL or the high rotational speed region correction amount IGL is selected by the selection means 54 depending on whether the rotational speed is in the high rotational speed region or not and outputted to the ignition timing calculation means 38.

Next, the operation of the ignition timing learning control system having such a configuration will be described using the flowcharts shown in FIGS. 3(a) to 3(e).

First, in steps 5too to 5103 of the routine shown in FIG. 3(a), the engine speed Ne and the intake pipe pressure Pm are determined.

As the water temperature Tw is read, it is determined whether or not knocking has occurred, and in step 9104.5105, the advance limit MBT and the basic ignition timing IGB are determined from the advance limit map and the basic ignition timing map based on the engine speed Ne and the intake pipe pressure Pm.
is searched. and steps 5106 to 5108
The learning execution conditions are checked and the knocking detection signal is on the high rotation speed side where there is a lot of disturbance, the low load side where the sensor output is small,
The learning condition is satisfied in any operating state other than the cold state, and the learning value update routine of step 5109 is executed.

That is, when starting the engine, first, the entire learning is selected in step 5200 of the routine of FIG. 3(b), and
In step 82, the address of the overall learning value AT is stored in the index register X. The presence or absence of knocking is determined in step 5203, and if knocking is present, the overall learning value AT is retarded by a certain amount γ in step 5204, the timer], 2 is cleared in step 5205, and step 8
At 20G, the counter is incremented to count the number of knocks. If there is no knocking, step 5
207, advance angle limit MBT and learning correction amount IGL (-AT
+AP), and if it falls within the region of MBT≦IGL, the learning value is not updated because even if the angle is advanced further, the advance limit MBT will be exceeded and the output torque will decrease. Then, in step 8208, timer 1, which measures the time without knocking, is checked, and if no knocking has occurred for a certain period of time 11 (for example, 1 second), and if the knocking feedback correction amount AK = 0 in step 5217, the overall learning value AT is determined in step 5209. is updated to advance by a certain amount a, and timer 1 is cleared in step 5210.

After step 8211, the overall learning termination condition is checked, and in step 5212, the magnitude of the overall learning value AT with respect to the predetermined maximum advance amount AM is determined, and if AT<AM, timer 2 is cleared in step 821B, In step 5214, the number of knocking is checked and the set value α
(for example, 5 times) or more, it is determined that the overall learning value AT has converged to the knocking limit, and the overall learning is terminated, and the end flag is set in step 5215. Further, when the advance angle progresses without knocking and reaches AT≧AM, steps 5212 to 52
Proceed to step 13 and set the cumulative time of timer 2 to 12 (for example, 3 seconds)
After that, complete the overall learning in the same way.

When the overall learning value AT is learned and determined in this way, the process proceeds from step 5200 to step 5202, where the address where the partial learning value AP of the current driving state is stored is entered into the index register is executed. That is, in the case of knocking, the partial learning value AP is updated to the retard side, and in the case of no knocking, the partial learning value AP is updated to the advance side under the condition of MBT>IGL. The learning update of the value AP is performed during operation as long as the learning condition is satisfied.

As a result, the learning correction amount IGL, which is the sum of the above-mentioned overall learning value AT and partial learning value AP, becomes a value very close to the actual knocking limit in each driving state.

On the other hand, after setting the knocking feedback correction amount AK in step 5I13, in step 911.0, the learning correction amount I is determined by the sum of the overall learning value AT and the partial learning value AP.
The value obtained by adding the knocking feedback correction amount AK to GL is compared with the advance limit MBT.

If MBT≦I GL+AK, step 5111 sets the ignition timing I based on the advance limit MBT and the basic ignition timing IGB.
GT is calculated. Therefore, the ignition timing (GT) is determined by the advance limit MBT characteristic as shown by the thick solid line in FIG.

Also, if MBT>IGL, step 5112
IGT-IGB+IGL+AK. If it is AK-0, the ignition timing GT is calculated from the learning correction amount IGL and the basic ignition timing IGB, and therefore the ignition timing IGT is parallel to the basic ignition timing IGB as shown by the thick solid line in Fig. 4. This results in a value closely aligned with the actual knocking limit IGT' for the octane rating.

Next, regarding the case where knocking occurs, the routine shown in FIG. 3(C) is executed in step 8113 to perform knocking feedback control.

That is, the process proceeds from step 5300 to 5304, where the knocking feedback correction amount AK is updated to a larger retard side by the retard amount γ, and in step 5305, timer 3 is cleared. At this time, the learning correction amount IGL is determined by steps 8208 and 3204 of the learning control routine shown in FIG. 3(b).
The ignition timing IGT in step 5118 of the routine of FIG. 3(a) is retarded by the retard amounts γ and γF. After that, when knocking is avoided, the process proceeds from step 5300 to step 5301 of the knocking feedback correction amount setting routine shown in FIG. In this case, the knocking feedback correction amount AK is determined in step 5302.
is gradually updated to the advance angle side by the advance angle amount a, and timer 3 is cleared in step 5303. Also in this case,
Steps 5217 to 521 of the routine of FIG. 3(b)
1, the partial learning value AP of the learning correction amount IGL is prohibited from advancing. In this way, when knocking occurs,
Learning control and knocking feedback control greatly retard the angle to prevent second and subsequent knocking from occurring in succession. After that, without knocking, the knocking feedback correction amount AK is first updated to the advance side until it becomes zero, and then the learning correction amount IGL can be updated to the advance side, and this two-way control allows mutual interference and learning. Variations in the correction amount IGL are prevented.

By the way, in a high rotational speed region of N e > 500 rpm, the learning condition in step 510B is not satisfied, and the learning value update in step 5109 is interrupted. and step 81
From 15 to 5tie, proceed to 5L17 and total learning value AT
The correction amount AL is estimated from the value APh on the highest rotation speed side of the partial learning value AP corresponding to the current suction pipe pressure Pm. Further, after setting the knocking feedback correction amount AK in step 5121, the process proceeds to step 8118, where the value obtained by adding the correction amount AL and the knocking feedback correction amount AK is compared with the advance limit MBT, and the smaller value is set at the high rotation speed. By adding the range correction amount to the basic ignition timing IGB in step 5L19.8120, the ignition timing I
GL is calculated.

As a result, the ignition timing IGT on the high rotation speed side becomes the same as the value set on the highest rotation speed side for each suction pipe pressure Pm, and since the operating conditions are similar, the occurrence of knocking is small and the ignition timing is sufficiently advanced. will be done.

On the other hand, if knocking occurs even in such a high rotational speed range, the routine shown in FIG. 3(e) is executed in step 5121, and the ignition timing IG is
'T is retarded. In knocking feedback control in this case, step 8306 is different from learning control.
Through steps 8308 to 8308, the knocking feedback correction amount AK is guarded by the set value AKo and retarded control is performed.

Above, we have described correcting the ignition timing in the high rotation speed range using the overall learning value of the learning control and the selected value of the partial learning value whose operating condition is similar to the partial learning value, but there are other methods as well. It is possible.

That is, the advance angle amount is ensured to a certain extent by the overall learned value AT of the learning control, and in addition to this, the minimum value of the partial learned values AP may be selected and used in order to suppress the occurrence of knocking. Furthermore, the overall learning value AT and partial learning value AP by these methods
The value may be set to the retard side by subtracting or multiplying the value by a predetermined value.

= 22 - [Effects of the Invention] As described above, according to the present invention, in the ignition timing learning control of a vehicle engine, in the high rotational speed region where knocking cannot be detected, the overall learning value and the partial learning value of the learning control are Since the ignition timing is estimated by determining the optimum correction amount using the learned value, the reliability of the correction amount is high, and the occurrence of knocking can be suppressed and the ignition timing can be sufficiently advanced to improve output and fuel efficiency.

Furthermore, by using both the overall learning value and the partial learning value of learning control, the degree of freedom and appropriateness in estimating the correction amount is increased.
It is preferable that the control form is also the same.

Furthermore, in knocking feedback control in a high rotation speed range, the knocking feedback correction amount is guarded, so even if the correction amount is incorrect due to mechanical noise, it will not become too large, and malfunctions can be prevented.

Further, by adding the partial learning value on the highest rotational speed side with similar driving conditions to the overall learning value, it is possible to suppress knocking and advance the engine angle to the maximum extent possible.

Furthermore, when the minimum value of the partial learning values is added to the overall learning value, the effect of suppressing the occurrence of knocking is high.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the ignition timing learning control method of the present invention, FIG. 2 is a block diagram showing an outline of a vehicle engine to which the present invention is applied, and FIG. 3 (a) to C) is a flowchart showing the action of ignition timing learning control, FIG. 4 is a characteristic diagram of the map, and FIG. 5 is a diagram showing a map of partial learning values. 1... Engine body, 7... Spark plug, 30...
- Control unit, 35... Basic ignition timing search means, 3
6... Lead angle limit search means, 37. Area discrimination means, 38
... Ignition timing calculation means, 40 ... Learning condition determination means, 41 ... Learning value updating means, 45 ... Partial learning value storage means, 46 ... Learning correction amount calculation means, 50 ... Selection means, 52... High rotation speed range correction amount setting means, 53.
- High rotation speed range determination means, 54... selection means.

Claims (3)

[Claims] (1) Execute overall learning to set the overall learning value while advancing or retarding depending on whether or not knocking occurs, and then setting partial learning values by advancing or retarding the angle depending on whether or not knocking occurs for each driving state. In this method, the ignition timing is corrected and determined based on the sum of the overall learning value and the partial learning value, and in the high engine speed range, the above overall learning and partial learning are interrupted, and the above overall learning value is An ignition timing learning control method characterized by calculating a high rotation speed range correction amount based on a learned value and a partial learning value, and determining the ignition timing in the high rotation speed range by correcting it with the high rotation speed range correction amount. (2) The ignition timing learning control method according to claim 1, wherein the high rotational speed range correction amount is calculated by adding a partial learning value on the highest rotational speed side for each engine load to at least the overall learning value. (3) Claim (1) wherein the high rotational speed range correction amount is calculated by adding at least the minimum value of the partial learning values to the overall learning value.
The ignition timing learning control method described.