JP5299355B2 - Ignition timing control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an ignition timing control device capable of properly performing learning of knock controlled variables. <P>SOLUTION: The ignition timing control device learns a deposit learning value (adep), and a second KCS learning value (agknkdp), as values for compensating steady state deviation offset between knock controlled variables updated based on occurrence situations of knocking and their reference values. A part of both a ratio learning region that learns the deposit learning value (adep), and a multipoint learning region that learns the second KCS learning value (agknkdp), are brought to be overlapped. In the ratio learning region, the relationship between an engine operating condition in the whole of the region and the deposit learning value (adep) is updated collectively based on a knock controlled variable (indicated by an arrow D in the figure). In the multipoint learning region, a plurality of learning regions are partitioned according to the engine operating condition while the second KCS learning values (agknkdp) are respectively set for respective partitions. In the multipoint learning region, the second KCS learning value (agknkdp) of a learning region including the engine operating condition at the time is updated based on knock controlled variables (indicated by arrows E1 to E3 in the figure). <P>COPYRIGHT: (C)2012,JPO&amp;INPIT

Description

  The present invention relates to an ignition timing control device that adjusts the ignition timing of an internal combustion engine in accordance with the occurrence of knocking.

  In general, in an internal combustion engine, knocking is detected by a knock sensor, and knock control is performed to adjust the ignition timing in accordance with the detection result. In this knock control, a control value (knock control amount) is set such that when the occurrence frequency of knocking is high, the ignition timing is retarded while when the occurrence frequency is low, the ignition timing is advanced. Then, the occurrence of knocking is suppressed by setting a control target value (required ignition timing) of the ignition timing based on this knock control amount.

  Here, the ignition timing at which the occurrence of knocking can be appropriately suppressed varies depending on individual differences, changes with time, operating environment, and the like of the internal combustion engine. Moreover, the degree of change in the ignition timing varies depending on the operating region of the internal combustion engine. For this reason, in a device in which knock control is executed, the amount of knock control may become a value that does not match the actual state as the engine operating state changes. Although this is a temporary phenomenon, it is not preferable because the accuracy of ignition timing control by knock control is reduced.

  Therefore, conventionally, in order to suppress such a decrease in ignition timing control accuracy, a device has been proposed that learns the relationship between the engine operating state and the knock control amount (specifically, the learning value) suitable for the engine operating state ( For example, see Patent Document 1). In this apparatus, the relationship between the engine operating state and the learned value is determined, and the change tendency of the relationship in the learning region is obtained and grasped in advance. Then, based on the change tendency, the above relationship in the entire learning region is collectively updated based on the knock control amount at that time.

  According to the apparatus, the relationship is updated in accordance with a change tendency expected in advance. Therefore, by calculating the learning value from the above relationship based on the engine operating state at that time and setting the required ignition timing based on the learning value, the amount of change in the knock control amount accompanying the change in the engine operating state is determined by the learning value. Compensation is made in advance, and a time commensurate with the actual situation is set as the required ignition timing.

JP 2005-147112 A

  Here, in the apparatus described in Patent Document 1, since the change tendency of the relationship between the engine operating state and the learned value is obtained and stored in advance, all of the stored change tendency and the actual change tendency are stored. It is difficult to match over the operating range of For this reason, it is inevitable that a difference between the stored change tendency and the actual change tendency occurs due to a setting error of the change tendency or individual differences among the internal combustion engines. For this reason, when the learning value is calculated based on the learned relationship as described above, an error occurs between the calculated value and a value in accordance with the actual condition, which causes a decrease in ignition timing adjustment accuracy. End up.

  In addition, if the knock control amount changes significantly due to disturbance such as noise superimposition on the detection signal of the knock sensor, not only the learning value corresponding to the engine operation area at this time but also learning over the entire learning area. The value is updated unnecessarily, which also causes a decrease in ignition timing adjustment accuracy.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide an ignition timing control device that can appropriately perform the learning of the knock control amount.

Hereinafter, means for achieving the above-described object and its operation and effects will be described.
The invention according to claim 1 updates the knock control amount in accordance with the occurrence of knocking, and learns a value for compensating for a steady deviation between the knock control amount and its reference value as a knock learning value. An ignition timing control device that sets a control target value for the engine ignition timing based on the knock control amount and the knock learning value, wherein the knock learning value includes a first knock learning value and a second knock learning value. And at least a part of the first execution area for learning the first knock learning value and the second execution area for learning the second knock learning value overlap, and the first execution area Based on the change tendency of the relationship between the engine operating state and the first knock learning value with respect to the determined change in the knock control amount, the relationship in the entire first execution region is collectively based on the knock control amount. The second execution region is divided according to the engine operation state, and a plurality of regions in which the second knock learning values are set separately are set in advance, and the engine operation at that time The gist is that the second knock learning value of the region including the state is updated based on the knock control amount.

  According to the above configuration, by learning in the first execution region, the relationship between the engine operating state and the first knock learning value is updated all over the entire first execution region based on the knock control amount at that time. Can do. Therefore, the above relationship, in other words, the first knock learning value can be quickly learned over a wide engine operation region with a small execution opportunity. The first knock learning value can reduce the steady deviation between the knock control amount and the reference value evenly over the entire first execution region.

  Moreover, the second knock learning in a plurality of regions is performed such that the second knock learning value in the region including the engine operating state at that time is updated based on the knock control amount at that time by learning in the second learning region. You can learn the values individually according to the actual situation. Therefore, the steady deviation between the knock control amount and its reference value can be accurately compensated for each of the plurality of regions by the second knock learning value.

Thus, according to the above configuration, the learning of the first knock learning value and the second knock learning value can be appropriately executed so that the deviation is appropriately reduced.
The learning of the first knock learning value and the second knock learning value and the setting of the control target value are the first in the engine operation region where the first execution region and the second execution region overlap, as in claim 2. The first knock learning value and the second knock learning value and the second knock learning value are both learned and used to set the control target value, and in the engine operation region where the first execution region and the second execution region do not overlap. Of the learning values, only the value corresponding to the region including the engine operating state at that time is learned and used for setting the control target value.

  According to such a configuration, the learning of the first knock learning value and the second knock learning value can be executed at appropriate timing, and the control target value of the ignition timing can be appropriately set based on these knock learning values. it can.

  According to a third aspect of the present invention, in the ignition timing control device according to the first or second aspect, in the engine operation region where the first execution region and the second execution region overlap, the deviation is compensated. The gist of the present invention is to set the distribution ratio of the update to the first knock learning value and the second knock learning value based on the engine operating state.

  The amount of change in the deviation due to the difference in the characteristics of the internal combustion engine varies depending on the engine operating region. According to the above configuration, when learning the relationship in the engine operation region where the variation amount of the deviation is small, the update amount of the relationship with respect to the deviation is suppressed to be small, and the first knock learning value is unnecessarily changed. It is possible to update the relationship in a stable state while suppressing. In addition, when learning the relationship in an engine operation region where the amount of change in the deviation is large, the amount of update of the relationship with respect to the deviation is made relatively large so that the relationship can be quickly converted into a relationship that matches the actual situation. Can be followed. As described above, according to the above configuration, it is possible to appropriately perform distribution of the updated first knock learning value for compensating for the deviation in a manner corresponding to the variation amount of the deviation which varies depending on the engine operation region, The first knock learning value and the second knock learning value can be appropriately learned with a high degree of freedom.

  Deposits derived from unburned fuel, lubricating oil, and the like gradually adhere to the intake valves and pistons of the internal combustion engine as time passes. When the deposit amount increases, knocking is likely to occur due to a decrease in the substantial volume of the combustion chamber and an increase in in-cylinder pressure during combustion. Normally, when the engine ignition timing, and thus the deviation, changes due to such deposits, the amount of change is greatest in the engine operating range where the engine load is medium (medium load operating range) while the medium load is the same. The smaller the distance from the operating area, the smaller.

  In this regard, the invention according to claim 4 is the ignition timing control device according to claim 3, wherein the distribution ratio is an engine operating region in which the ratio of the distribution to the first knock learning value is medium engine load. The gist of this is that it is the largest value and the value that decreases as the distance from the region increases.

  According to this configuration, the distribution ratio can be set according to the relationship between the engine load and the variation amount of the deviation, and the first knock learning value and the second knock learning value can be learned appropriately. .

  Further, when the deviation changes due to the adhesion of the deposit, the amount of change becomes the largest in the engine operation range where the engine speed is medium (medium rotation operation region), while leaving the middle rotation operation region. It gets smaller.

  According to a fifth aspect of the present invention, in the ignition timing control device according to the third or fourth aspect, the distribution ratio is an engine operating region in which the ratio of the distribution to the first knock learning value is medium. The gist of this is that it is the largest value and the value that decreases as the distance from the region increases.

  According to such a configuration, the distribution ratio can be set according to the relationship between the engine rotation speed and the variation amount of the deviation, and the first knock learning value and the second knock learning value can be learned appropriately. .

  The invention according to claim 6 is the ignition timing control device according to any one of claims 1 to 5, wherein the knock learning value further includes a third knock learning value, and the third knock learning value. Is learned in an engine operation region that does not overlap with any of the first execution region and the second execution region, and is used for setting the control target value in all engine operation regions. To do.

  According to the above configuration, when there is a factor (for example, fuel property) having the same degree of influence on the steady deviation between the knock control amount and its reference value in all engine operating ranges, the phase of the factor The change in the deviation due to the difference can be compensated through the learning of the third knock learning value and the reflection in the control target value. In addition, learning of the first knock learning value, the second knock learning value, and the third knock learning value can be performed at appropriate timings, and the control target value of the ignition timing is appropriately set based on these knock learning values. Can be set.

  The invention according to claim 7 is the ignition timing control device according to claim 6, wherein the device is an engine operation region in which only the first execution region of the first execution region and the second execution region overlaps. The first knock learning value and the third knock learning value are learned, and the engine load is determined as a ratio for distributing the update for compensating for the deviation to the first knock learning value in the learning. The gist of the invention is to set a value that is the largest in the medium engine operation region and that decreases as the distance from the region increases.

  According to the above configuration, the update for compensating for the deviation is distributed to the first knock learning value and the third knock learning value in accordance with the relationship between the engine load and the variation of the deviation. The first knock learning value and the third knock learning value can be appropriately learned.

  The invention according to claim 8 is the ignition timing control device according to claim 6 or 7, wherein the device is an engine in which only the first execution region of the first execution region and the second execution region overlaps. In the operation region, the first knock learning value and the third knock learning value are learned, and in the learning, an update amount for compensating the deviation is distributed to the first knock learning value as an engine. The gist of the invention is to set a value that becomes the largest in the engine operation region where the rotational speed is medium and becomes smaller as the distance from the region increases.

  According to the above configuration, the update for compensating for the deviation is distributed to the first knock learning value and the third knock learning value in accordance with the relationship between the engine rotational speed and the variation of the deviation. And the first knock learning value and the third knock learning value can be appropriately learned.

BRIEF DESCRIPTION OF THE DRAWINGS Schematic which shows schematic structure of the ignition timing control apparatus concerning one Embodiment which actualized this invention. The schematic diagram which shows the setting aspect of ignition timing. Explanatory drawing which shows the execution area | region which performs learning of each knock learning value. The flowchart which shows the execution aspect of update and reflection of each knock learning value. The table | surface which shows the execution aspect of learning and reflection of each knock learning value in each execution pattern. Schematic diagram showing the relationship between engine operating state and distribution ratio. 6 is a schematic diagram illustrating an example of each knock learning value and a combined correction amount.

Hereinafter, an embodiment in which an ignition timing control device according to the present invention is embodied will be described.
FIG. 1 shows a schematic configuration of an ignition timing control device according to the present embodiment.
As shown in FIG. 1, the combustion chamber 13 of the internal combustion engine 10 is provided with a spark plug 14 that ignites and burns an air-fuel mixture sucked into the combustion chamber 13. The cylinder block 11 of the internal combustion engine 10 is provided with a knock sensor 15 for detecting the occurrence of knocking due to the combustion of the air-fuel mixture.

  Various controls relating to the operation of the internal combustion engine 10 are performed by the electronic control unit 16. The electronic control unit 16 includes a CPU that executes various controls, a memory that stores information necessary for the control, an input port for inputting a signal from the outside, an output port for outputting a command signal to the outside, and the like. It is prepared for.

  Various sensors for detecting the engine operating state are connected to the input port of the electronic control unit 16. As various sensors, in addition to the knock sensor 15, a crank angle sensor 17 for detecting a crank angle, which is a rotation phase of a crankshaft (not shown), and a throttle valve opening (throttle opening TA) are detected. A throttle sensor 19 and the like are connected to this input port. The detection signals of these sensors are input to the electronic control unit 16 through the input port. The engine rotational speed NE can also be obtained from the detection signal of the crank angle sensor 17.

  On the other hand, the output port of the electronic control unit 16 is connected to a drive circuit for actuators necessary for engine control, such as an igniter 14a for generating a high-voltage current necessary for ignition of the air-fuel mixture by the spark plug 14. The electronic control unit 16 performs various calculations based on the detection signals of the sensors, and performs engine control by executing drive control of actuators based on the calculation results.

  In the internal combustion engine 10 configured as described above, the electronic control unit 16 executes “knock control” for adjusting the ignition timing in accordance with the occurrence of knocking detected by the knock sensor 15. Hereinafter, an execution mode of the “knock control” will be described with reference to FIG.

  In the knock control according to the present embodiment, a required ignition timing afin that is a control target value of the ignition timing is set. Here, the ignition timing is expressed as the advance amount [° CA] of the crank angle with respect to the compression top dead center of the cylinder to be ignited.

  As shown in FIG. 2, when setting the required ignition timing afin, first, the most advanced ignition timing absef which is the limit value on the advance side of the setting range of the required ignition timing afin in the knock control, and the limit on the retard side thereof The most retarded ignition timing akmf, which is a value, is calculated. Based on these, the maximum retardation amount akmax of the required ignition timing afin with respect to the most advanced ignition timing absef during knock control is calculated.

  The most advanced ignition timing absef is calculated based on the MBT point ambt and the first knock limit point aknok1. Specifically, as shown in the following formula [1], a value on the more retarded side of the MBT point ambt and the first knock limit point aknok1 is set as the most advanced ignition timing absef.

absef = min (ambt, aknok1) [1]

The MBT point ambt indicates the ignition timing (maximum torque ignition timing) at which the maximum torque is obtained under the current engine operating conditions. Further, the first knock limit point aknok1 is an ignition timing advance limit value (knock value) that can be kept within a level where knocking can be allowed under the best possible conditions when using a high octane fuel with a high knock limit. Limit ignition timing). The MBT point ambt and the first knock limit point aknok1 are set with reference to the setting map stored in the memory of the electronic control unit 16 based on the current engine speed NE, the engine load KL, and the like. In the present embodiment, the engine load KL is calculated based on the intake air amount of the internal combustion engine 10. The engine load KL can also be calculated based on parameters other than the intake air amount, for example, the fuel injection amount, the throttle opening TA, the accelerator pedal operation amount, and the like.

  On the other hand, the most retarded ignition timing akmf is set as an index value of an ignition timing that can be kept within a level that sufficiently allows the occurrence of knocking even under worst-case conditions. Specifically, as the most retarded ignition timing akmf, as shown in the following equation [2], a value delayed by the deposit learning value adep from the second knock limit point aknok2 is set.

akmf = aknok2-adep ... [2]

Here, deposits derived from unburned fuel, lubricating oil, and the like gradually adhere to the intake valves and pistons (not shown) of the internal combustion engine 10 over time. When the deposit amount increases, the substantial volume of the combustion chamber 13 of the internal combustion engine 10 decreases, and the in-cylinder pressure at the time of combustion increases, so that knocking is likely to occur.

  The second knock limit aknok2 can be within an acceptable level of knocking when using a low octane fuel with a low knock limit under the best possible conditions, for example, when no deposit is deposited. It shows the advance limit (ignition limit ignition timing) of the ignition timing. The second knock limit point aknok2 is also set with reference to the setting map stored in the memory of the electronic control unit 16 based on the current engine speed NE, the engine load KL, and the like.

  The deposit learning value adep is set as a knock learning value for compensating for the change in the ignition timing due to the adhesion of such deposit. The deposit learning value adep is an index value indicating the retard amount of the ignition timing according to the deposit degree of deposit of the current internal combustion engine 10. Specifically, as shown in the following equation [3], a value obtained by multiplying the retardation width DLAKNOK by the ratio learning value rgknk is set as the deposit learning value adep.

adep = DLAKNOK × rgknk [3]

The retard width DLAKNOK indicates the retard amount of the required ignition timing afin due to the influence of deposit adhesion in a state where the assumed maximum amount of deposit adheres under a predetermined engine operating condition where the influence of deposit adhesion appears most prominently. Specifically, it is a value indicating the retard amount of the most retarded ignition timing akmf due to the effect of deposit adhesion. The retard width DLAKNOK is calculated by subtracting the third knock limit point aknok3 from the second knock limit point aknok2 as shown in the following equation [4].

DLAKNOK = aknok2-aknok3 [4]

The third knock limit point aknok3 is the occurrence of knocking when, for example, the deposit amount described above has reached the maximum amount assumed under the worst conditions when using a low octane fuel with a low knock limit. Represents the advance limit (ignition limit ignition timing) of the ignition timing that can fall within an allowable level. Note that the value of the third knock limit point aknok3 is set in consideration of the current engine speed NE, the engine load KL, and the like.

  The ratio learning value rgknk is set as an index value indicating the degree of deposit adhesion to the internal combustion engine 10. Here, a state where there is no deposit adherence is set to a value “0”, a state where the amount of deposit adherence is assumed to be a maximum value, a value “1”, and the degree of deposit adherence is a value of the ratio learning value rgknk. Is expressed as.

  The ratio learning value rgknk is set to “0” as an initial value at the time of shipment from the factory where no deposit is attached. Thereafter, the value of the ratio learning value rgknk is gradually increased or decreased within the range of “0” or more and “1” or less according to the occurrence frequency of knocking detected by the knock sensor 15. Specifically, if the occurrence frequency of knocking increases, the value of the ratio learning value rgknk is gradually increased, and if the occurrence frequency of knocking decreases, the value is gradually decreased. Details of the update mode of the ratio learning value rgknk will be described later.

  Thus, by updating the ratio learning value rgknk according to the occurrence frequency of knocking, a value (deposit learning value adep) obtained by multiplying the ratio learning value rgknk and the retardation angle DLAKNOK depends on the degree of deposit adhesion. Will be updated.

Based on the most advanced ignition timing absef and the most retarded ignition timing akmf, the maximum retard amount akmax is calculated from the following equation [5].

akmax = absef−akmf [5]

The required ignition timing afin is calculated by obtaining an ignition timing retard amount ank that is a retard amount of the required ignition timing afin with respect to the most advanced ignition timing absef. The ignition timing retardation amount ankk is set based on the maximum retardation amount akmax, the first KCS learning value agknk, the second KCS learning value agnkddp, and the KCS feedback correction value akcs, as shown in the following equation [6].

aknk = akmax-agknk-agknkdp + akcs [6]

Then, as shown in the following equation [7], the required ignition timing afin is set by subtracting the ignition timing retard amount ankk from the most advanced ignition timing absef. That is, as shown in the following equation [8], the required ignition timing afin is set with respect to the most retarded ignition timing akmf set by delaying the second knock limit point aknok2 by the deposit learning value adep. The first KCS learning value agknk and the second KCS learning value agnknkdp are advanced by a value, while the KCS feedback correction value akcs is a delayed value.

afin = absef−aknk [7]
afin = akmf + (agknk + agknkdp) −akcs [8]

Note that the value of the ignition timing retardation amount ankk is limited so that the required ignition timing afin does not reach the timing of the advance side of the most advanced ignition timing absef. For example, when the required ignition timing afin is calculated based on Equation [7], the ignition timing retardation amount ankk is limited to a value of “0” or more. Therefore, when the ignition timing retardation amount ankk calculated based on the equation [6] becomes a negative value, the value is set to “0”.

  The KCS feedback correction value akcs is set according to the occurrence of knocking detected by the knock sensor 15. Specifically, when it is determined that the detected knocking level is less than a predetermined determination value and the occurrence of knocking is within a sufficiently allowable level, the value of the KCS feedback correction value akcs is gradually decreased. . When the detected knocking level is equal to or higher than the determination value, the KCS feedback correction value akcs is gradually increased. In the present embodiment, this KCS feedback correction value akcs functions as a knock control amount.

  In the present embodiment, the first KCS learning value agknk, the second KCS learning value agknkdp, and the deposit learning are used as a knock learning value for compensating for a steady deviation between the KCS feedback correction value akcs and its reference value (= 0). Three values of value adep are set.

  These knock learning values reflect changes in ignition timing due to different factors. Specifically, the deposit learning value adep reflects the change in ignition timing due to the effect of deposit adhesion. The first KCS learning value agknk mainly reflects the change in ignition timing due to the influence of the fuel property (for example, octane number). Further, the second KCS learning value agknkdp mainly reflects changes in the ignition timing due to individual differences, changes with time of the internal combustion engine 10, and changes in the operating environment. In the present embodiment, deposit learning value adep functions as a first knock learning value, second KCS learning value agnkdp functions as a second knock learning value, and first KCS learning value agknk functions as a third knock learning value. . In addition, the ratio learning region that performs learning of the deposit learning value adep functions as a first execution region, and the multipoint learning region that performs learning of the second KCS learning value agnkdpp functions as a second execution region.

  Each knock learning value is gradually reduced when the absolute value of the KCS feedback correction value akcs is larger than the predetermined value (| akcs |> A) for a predetermined time or longer. To be updated. The update amount (learning update amount tdl) is calculated based on the KCS feedback correction value akcs. Specifically, the larger the absolute value of the KCS feedback correction value akcs is, the larger the learning update amount tdl is calculated.

  When the state where the KCS feedback correction value akcs is larger than the positive value [A] (akcs> A) continues, the learning update amount tdl is subtracted from any of the knock learning values. At the same time, the learning update amount tdl is also subtracted from the KCS feedback correction value aks. On the other hand, when the state where the KCS feedback correction value akcs is smaller than the negative value [−A] (akcs <−A) continues, the learning update amount tdl is added to any of the knock learning values and the KCS feedback correction value akcs, respectively. Is done.

The knock learning values updated in this way are stored in the backup memory of the electronic control unit 16 and are retained even when the engine is stopped.
Through such knock control, the required ignition timing afin is set to a value on the advance side that allows a large torque to be obtained within a range in which knocking exceeding an allowable level does not occur.

Hereinafter, the execution mode of the update of each knock learning value and the reflection to the required ignition timing afin will be described in detail with reference to FIGS.
FIG. 3 shows the relationship between execution regions in which learning of each knock learning value is executed. FIG. 4 shows an execution procedure for updating and reflecting each knock learning value. Further, FIG. 5 shows an execution mode of learning and reflecting each knock learning value in each execution pattern.

  In the present embodiment, four patterns of [Pattern 1] to [Pattern 4] selected based on the engine operating state are set as execution patterns for executing update and reflection of each knock learning value. Hereinafter, [Pattern 1] to [Pattern 4] will be described separately.

  [Pattern 1] is a multipoint learning region (region indicated by “B” in FIG. 3) and a ratio learning region (region indicated by “C” in FIG. 3) among all engine operation regions (basic operation regions). ) Is selected in an engine operation region that does not overlap with any of (No in steps S101 to S103 in FIG. 4).

  In [Pattern 1], as shown in FIG. 5, only the first KCS learning value agknk among the knock learning values is learned (step S104 in FIG. 4). Specifically, the learning update amount tdl is added to the first KCS learning value agknk. In calculating the required ignition timing afin, only the first KCS learning value agknk is reflected (step S105). At this time, both “0” are set as the second KCS learning value agnknkdp and the ratio learning value rgknk. That is, only the first KCS learning value agknk among the knock learning values is used for setting the required ignition timing afin.

  The degree of influence of the difference in the octane number of the fuel on the steady deviation between the KCS feedback correction value akcs and its reference value is almost the same in all engine operating ranges. The engine operation region in which [Pattern 1] is selected is a region in which the variation in ignition timing due to the difference in the octane number of such fuel appears in the deviation. Therefore, in the engine operation region, the change in the deviation is reflected in the first KCS learned value agknk. Therefore, by setting the required ignition timing afin based on the first KCS learned value agknk, the change in the ignition timing due to the difference in the octane number of the fuel is compensated.

  [Pattern 2] is an engine operation region in which the ratio learning region (region indicated by “C” in FIG. 3) does not overlap in the basic operation region, and the basic operation region and the multipoint learning region (in FIG. 3). Are selected in the engine operation region (step S101: NO and step S102: YES in FIG. 4).

  In [Pattern 2], as shown in FIG. 5, only the second KCS learning value agnkddp of each knock learning value is learned (step S106 in FIG. 4). Specifically, the learning update amount tdl is added to the second KCS learning value agnknkdp. Further, when calculating the required ignition timing afin, the first KCS learning value agknk and the second KCS learning value agknkdp are reflected (step S107). At this time, “0” is set as the deposit learning value adep. That is, only the first KCS learning value agknk and the second KCS learning value agnknkdp among the knock learning values are used for setting the required ignition timing afin.

  In the multi-point learning region, as indicated by broken lines in FIG. 3, a plurality of (second in the present embodiment, nine second KCS learning values agnkdp) that are divided according to the engine speed NE and the engine load KL are set. Learning areas) are preset. Then, in [Pattern 2], the learning update amount tdl is added to the second KCS learning value agnkdp in the learning region including the engine operating state determined by the engine load KL and the engine speed NE at that time. Learning of the 2KCS learning value agnknkdp is executed.

  Depending on the engine operating region, variations occur in the degree of influence of individual differences, changes over time, and changes in the operating environment of the internal combustion engine 10 on the steady deviation between the KCS feedback correction value akcs and its reference value. As the multi-point learning region, an engine operation region in which such a variation of the steady deviation appears is set. Therefore, in this multipoint learning region, the second KCS learning value agnkdp in the learning region including the engine operating state at that time is updated based on the KCS feedback correction value akcs at that time. The 2KCS learning value agknkdp is learned separately in accordance with the actual steady deviation. Therefore, by setting the required ignition timing afin based on the second KCS learning value agnknkdp, the deviation is accurately compensated for each of the plurality of learning regions by the second KCS learning value agnkkdp.

  [Pattern 3] is an engine operation region in which the multipoint learning region (region indicated by “B” in FIG. 3) does not overlap in the basic operation region, and the basic operation region and the ratio learning region (FIG. 3). (The region indicated by “C” in FIG. 4) is selected (both “NO” in steps S101 and S102 of FIG. 4 and YES in step S103).

  In [Pattern 3], as shown in FIG. 5, the first KCS learning value agknk and the deposit learning value adep (specifically, the ratio learning value rgknk) among the knock learning values are learned. Specifically, the distribution ratio tk is calculated based on the engine speed NE and the engine load KL (step S108 in FIG. 4), and the learning update amount tdl is calculated based on the distribution ratio tk to the first KCS learning value. It is distributed and added to agknk and the ratio learning value rgknk (step S109). The calculation mode of the distribution ratio tk and the distribution method of the learning update amount tdl in [Pattern 3] will be described later. Further, when calculating the required ignition timing afin, the first KCS learning value agknk and the deposit learning value adep are reflected. At this time, “0” is set as the second KCS learning value aggnkdp. That is, only the first KCS learning value agknk and the deposit learning value adep among the knock learning values are used for setting the required ignition timing afin.

  In the ratio learning area, the relationship between the engine operating state determined by the engine speed NE and the engine load KL and the deposit learning value adep is learned and stored. In the present embodiment, a change tendency of the above relationship with respect to a steady change between the KCS feedback correction value akcs and its reference value based on an experimental result, a simulation result, etc., and a change in ignition timing caused by deposit adhesion The change tendency of the above relationship that can suppress the minute is obtained in advance, and the above relationship is set based on the change tendency. Specifically, the setting map of the second knock limit point aknok2, the setting map of the third knock limit point aknok3, and the above equation [3] are set as this relationship.

  Therefore, in this ratio learning area, the above-described relationship (specifically, the ratio learning value rgknk) is updated based on the KCS feedback correction value akcs as described above, so that the deposit learning value adep in the entire ratio learning area is Based on the occasional KCS feedback correction value aks, it is updated all at once. Therefore, it is quickly learned over a wide engine operation region at an execution opportunity with a small deposit learning value adep. Then, by such deposit learning value adep, the steady deviation between the KCS feedback correction value akcs and its reference value is reduced uniformly throughout the ratio learning region.

  [Pattern 4] is an engine operation in which the basic operation region, the multipoint learning region (region indicated by “B” in FIG. 3), and the ratio learning region (region indicated by “C” in FIG. 3) all overlap. The area is selected (step S101 in FIG. 4: YES).

  In [Pattern 4], as shown in FIG. 5, the second KCS learning value agnknkdp and the deposit learning value adep among the knock learning values are learned. Specifically, the distribution ratio tk is calculated based on the engine speed NE and the engine load KL (step S111 in FIG. 4), and the learning update amount tdl is calculated based on the distribution ratio tk to the second KCS learning value. It is distributed and added to agknkdp and the ratio learning value rgknk (step S112). The calculation mode of the distribution ratio tk and the distribution method of the learning update amount tdl in [Pattern 4] will be described later. Further, in calculating the required ignition timing afin, all knock learning values (first KCS learning value agknk, second KCS learning value agknkp, and deposit learning value adep are reflected (step S113). Used to set the required ignition timing afin.

  As described above, when updating the knock learning value in [Pattern 3] or [Pattern 4], the distribution is made to the deposit learning value adep and other knock learning values based on the engine speed NE and the engine load KL. A distribution ratio tk is set. This distribution ratio tk indicates the ratio of the amount reflected in the ratio learning value rgknk in the learning update amount tdl. Specifically, as the distribution ratio tk, when all of the learning update amount tdl is reflected in the ratio learning value rgknk, “1” is set, while all of the learning update amount tdl is other than the ratio learning value rgknk. When reflecting the knock learning value, a value within a range of “0” to “1”, such as “0”, is set.

  When deposit adheres to the internal combustion engine 10, knocking is likely to occur, and the steady deviation between the KCS feedback correction value akcs and its reference value changes. The steady deviation variation varies depending on the operating region of the internal combustion engine 10.

  Based on this point, in the present embodiment, the distribution ratio tk in [Pattern 3] and the distribution ratio tk in [Pattern 4] are set based on the engine rotational speed NE and the engine load KL. This makes it possible to set the distribution ratio tk according to the amount of change in the deviation due to deposit adhesion.

  Therefore, when the relationship between the deposit learning value adep and the engine operating state is learned in the engine operation region where the change amount of the deviation is small, by setting a small value as the distribution ratio tk, the learning update amount tdl is set. The update amount of the above relationship (specifically, the ratio learning value rgknk) can be kept small. Therefore, in this case, the above relationship is updated in a stable updating manner while suppressing unnecessary fluctuations in the deposit learning value adep.

  In addition, when the relationship is learned in the engine operation region where the variation amount of the deviation is large, an update amount of the ratio learning value rgknk with respect to the learning update amount tdl is set by setting a relatively large value as the distribution ratio tk. Is relatively large. Therefore, in this case, the relationship immediately follows the relationship that matches the actual situation.

  As described above, according to the present embodiment, the learning update amount tdl is distributed to the deposit learning value adep and other knock learning values in accordance with the amount of change in the deviation, which differs depending on the operating region of the internal combustion engine 10. Can do. Therefore, when [Pattern 3] is selected, the first KCS learning value agknk and the second KCS learning value agknkdp can be appropriately learned with a high degree of freedom. When [Pattern 4] is selected, the second KCS learning value agnknkdp and the deposit learning value adep can be appropriately learned with a high degree of freedom.

  Here, normally, when the ignition timing, and thus the above-mentioned steady deviation, changes due to deposits on the internal combustion engine 10, the amount of change is medium and the engine speed NE is medium. It becomes the largest in a certain operation region (medium load / medium rotation operation region), while it becomes smaller as the distance from the medium load / medium rotation operation region increases. The reason will be described below.

  In the internal combustion engine 10, when the ignition timing is changed by a predetermined amount, the combustion state in the operating region where the combustion state of the fuel in the combustion chamber 13 is good changes relatively greatly, whereas the operating region where the combustion state is poor The characteristic is that the amount of change in the combustion state is small.

  In the engine operation region where the engine load KL is large (high load operation region) and the engine operation region where the engine rotational speed NE is high (high rotation operation region), since securing of output torque is important, the fuel combustion state is Engine control is executed so as to improve the engine. On the other hand, in the medium load operation region and the medium rotation operation region, since the emphasis is on the suppression of the fuel consumption, the engine control is executed so as to make the combustion state slightly worse in order to realize this. Therefore, when the ignition timing is retarded through knock control to suppress knocking caused by deposit adhesion, the combustion state is changed with a small retardation amount in the high load operation region and the high rotation operation region where the combustion state is good. In contrast, the occurrence of knocking is suppressed, whereas in the middle load operation region and the intermediate rotation operation region where the combustion state is relatively poor, the occurrence of knocking cannot be suppressed to an appropriate level unless the retard amount is increased. For this reason, in the operation region where the engine load KL is large or the operation region where the engine rotational speed NE is high, the steady state due to deposit adhesion to the internal combustion engine 10 as the engine load KL decreases or the engine rotational speed NE decreases. It can be said that the degree of change of a large deviation becomes large.

  On the other hand, in the operation region where the engine load KL is small, the amount of heat generated in the combustion chamber 13 decreases as the engine load KL decreases, so knocking is less likely to occur. Further, in the operation region where the engine rotational speed NE is low, the lower the engine rotational speed NE is, the longer the fuel combustion interval in the combustion chamber 13 becomes, so that knocking is less likely to occur. Therefore, the smaller the engine load KL is in the operating region where the engine load KL is smaller, and the lower the engine rotational speed NE is in the operating region where the engine rotational speed NE is lower, the more the ignition timing due to deposit adhesion to the internal combustion engine 10 becomes. It can be said that the amount of change becomes smaller, and the degree of change in the steady deviation becomes smaller.

For this reason, in the apparatus according to the present embodiment, the steady deviation change degree becomes the largest in the medium load / middle load operation region.
In the present embodiment based on such circumstances, as shown in FIG. 6, the distribution ratio tk is set to the largest value in the middle rotation load operating region (the region indicated by “TK1” in the drawing). In other words, the smaller the value is set as the distance from the region TK1 is, that is, the distribution ratio tk gradually decreases in the order of the middle rotation load region TK1, the region TK2, the region TK3, and the region TK4. The ratio tk is set with reference to a setting map stored in the memory of the electronic control unit 16 based on the current engine speed NE, engine load KL, and the like.

  Based on the distribution ratio tk, the update amount Δagk of the first KCS learning value agknk (or the second KCS learning value agknkdp) is calculated from the following equation [9], and the ratio learning value rgknk is updated from the following equation [10]. The quantity Δrgk is calculated.

Δagk = tk × tdl ... [9]
Δrgk = (1−tk) × (−tdl) ÷ DLAKNOK [10]

In the update of the knock learning value in [Pattern 3], the update amount Δagk is added to the first KCS learning value agknk, and the update amount Δrgk is added to the ratio learning value rgknk. On the other hand, in the update of the knock learning value in [Pattern 4], the update amount Δagk is added to the second KCS learning value agknkdp, and the update amount Δrgk is added to the ratio learning value rgknk.

  As described above, in the present embodiment, the distribution ratio tk is set in accordance with the relationship between the engine load KL and the steady variation of the deviation and the relationship between the engine speed NE and the variation of the deviation. The learning update amount tdl can be distributed to the deposit learning value adep and other knock learning values based on the distribution ratio tk. Therefore, it is possible to appropriately learn the deposit learning value adep and the other knock learning values.

Hereinafter, the operation of learning each knock learning value in this way will be described.
FIG. 7 shows the knock learning values and the combined correction amount based on the knock learning values.
In the apparatus according to the present embodiment, through learning of the deposit learning value adep in the ratio learning region, the relationship between the engine operating state and the deposit learning value adep is determined based on the KCS feedback correction value akcs at that time. (See arrow D in FIG. 7). Therefore, the deposit learning value adep can be quickly learned over a wide engine operation region with a small execution opportunity. The deposit learning value adep can reduce the steady deviation between the KCS feedback correction value akcs and its reference value evenly over the entire ratio learning region.

  Here, in the ratio learning area, since the change tendency of the relationship between the engine operation state and the deposit learning value adep is obtained and stored in advance, the stored change tendency and the actual change tendency are all stored in the engine operation state. It is difficult to match across regions. For this reason, the relationship between the learned deposit learning value adep and the engine operating state may deviate from the relationship suitable for the actual change tendency due to such a setting error of the change tendency or individual differences among the internal combustion engines 10. Unavoidable. Therefore, when the deposit learning value adep is learned as described above, an error occurs between the deposit learning value adep and a value in accordance with the actual situation, which causes a decrease in ignition timing adjustment accuracy. End up. In addition, if the KCS feedback correction value akcs changes significantly due to disturbance such as noise superposition on the detection signal of the knock sensor 15, not only the deposit learning value adep corresponding to the engine operating area at this time but also the ratio The deposit learning value adep is unnecessarily updated over the entire learning region. This also causes a decrease in ignition timing adjustment accuracy.

  In this regard, in this embodiment, the second KCS learning value agnkdp in the learning region including the engine operating state at that time is updated based on the KCS feedback correction value akcs at that time through learning in the multipoint learning region. In addition, the second KCS learning values agnkdp of a plurality of learning areas can be learned individually in a manner that is in accordance with the actual situation. Therefore, the second KCS learning value aggnkdp is used for each of a plurality of learning regions (see arrows E1, E2, and E3 in FIG. 7), and the steady deviation between the KCS feedback correction value akcs and its reference value is accurately compensated. Will be able to. As described above, according to the present embodiment, the learning of the deposit learning value adep and the learning of the second KCS learning value agnkdp can be appropriately executed so that the deviation is appropriately reduced.

  By the way, if the configuration for performing the learning of the deposit learning value adep in the ratio learning region is omitted, in order to appropriately suppress the deviation in all engine operation regions, all of the plurality of learning regions in the same multipoint learning region are used. Since it is necessary to secure the execution frequency of learning, the time required for the learning becomes long.

  In this regard, in the apparatus according to the present embodiment, the deviation can be quickly reduced over the entire ratio learning region through learning of the deposit learning value adep in the ratio learning region. In addition, the deviation can be reduced in each learning region through learning of the second KCS learning value agnkddp in the multipoint learning region. Therefore, the deviation can be quickly and accurately suppressed over the entire engine operation region as compared with a device that omits the configuration for performing learning of the deposit learning value adep in the ratio learning region.

  Furthermore, the first KCS learning value agknk is learned in the engine operation region in which [Pattern 1] or [Pattern 3] is selected (see arrow F in FIG. 7), and the first KCS learning is performed in all engine operation regions. The required ignition timing afin is set based on the value agknk. Therefore, the first KCS learned value agknk can accurately compensate for the change in ignition timing due to the difference in the octane number of the fuel over the entire engine operation range.

  Thus, according to the present embodiment, each knock learning value can be appropriately learned and updated. When [Pattern 1] is selected, the required ignition timing afin is corrected to the advance side by the amount corresponding to the first KCS learning value agknk. When [Pattern 2] is selected, the required ignition timing afin is corrected to the advance side by an amount corresponding to the addition of the first KCS learning value agnknk and the second KCS learning value agnkdp. Further, when [Pattern 3] is selected, the angle is advanced by an amount corresponding to the first KCS learning value agknk and retarded by an amount corresponding to the deposit learning value adep (the amount of the combined correction amount indicated by the line L1 in the figure). Only the required ignition timing afin is corrected. When [Pattern 4] is selected, the angle is advanced by an amount corresponding to the sum of the first KCS learning value agknk and the second KCS learning value aggnkdp, and is retarded by an amount corresponding to the deposit learning value adep ( The required ignition timing afin is corrected by the combined correction amount indicated by the line L2 in the drawing.

As a result, an appropriate value is set as the required ignition timing afin based on each knock learning value in accordance with the engine operating state.
As described above, according to the present embodiment, the effects described below can be obtained.

  (1) Through learning the deposit learning value adep in the ratio learning region, the relationship between the engine operating state and the deposit learning value adep is updated all over the entire ratio learning region based on the KCS feedback correction value aks. be able to. Therefore, the deposit learning value adep can be quickly learned over a wide engine operation region with a small execution opportunity. The deposit learning value adep can reduce the steady deviation between the KCS feedback correction value akcs and its reference value evenly over the entire ratio learning region. In addition, through learning in the multipoint learning region, the second KCS learning values agnkdp in the plurality of learning regions can be learned individually in a form that is in accordance with the actual situation. Therefore, the steady deviation between the KCS feedback correction value akcs and its reference value can be accurately compensated for each of the plurality of learning regions by the second KCS learning value aggnkdp. Therefore, according to the present embodiment, the learning of the deposit learning value adep and the learning of the second KCS learning value agnkdp can be appropriately executed so that the deviation is appropriately reduced.

  (2) When selecting [Pattern 4], the deposit learning value adep and the second KCS learning value agnknkdp are both learned and used to set the required ignition timing afin. When [Pattern 2] is selected, only the second KCS learning value agnkddp of the deposit learning value adep and the second KCS learning value agnknkdp is learned and used for setting the required ignition timing afin. Further, when [Pattern 3] is selected, only the deposit learning value adep of the deposit learning value adep and the second KCS learning value aggnkdp is learned and used for setting the required ignition timing afin. Therefore, the learning of the deposit learning value adep and the second KCS learning value agknkdp can be executed at appropriate timing, and the required ignition timing afin can be appropriately set based on the deposit learning value adep and the second KCS learning value agknkdp. it can.

  (3) The distribution ratio tk when [Pattern 4] is selected is set based on the engine speed NE and the engine load KL. Therefore, the learning update amount tdl can be distributed between the deposit learning value adep and the second KCS learning value agnkdpp in accordance with the steady deviation change amount that varies depending on the operating region of the internal combustion engine 10, and these deposit learning values. It is possible to appropriately learn the adep and the second KCS learning value aggnkdp with a high degree of freedom.

  (4) As the distribution ratio tk at the time of selecting [Pattern 4], the largest value is set in the medium load operation region, and the smaller the value is set as the distance from the medium load operation region is. Therefore, the distribution ratio tk can be set in accordance with the relationship between the engine load KL and the amount of change in the steady deviation, and the learning update amount tdl is determined as the deposit learning value adep based on the distribution ratio tk. And the second KCS learning value agnknkdp can be appropriately distributed.

  (5) As the distribution ratio tk at the time of selecting [Pattern 4], the largest value is set in the middle rotation operation region, and the smaller the value is set as the distance from the middle rotation operation region is. Therefore, the distribution ratio tk can be set in accordance with the relationship between the engine rotational speed NE and the amount of change in the steady deviation, and the learning update amount tdl is set as the deposit learning value based on the distribution ratio tk. It is possible to appropriately distribute the adep and the second KCS learning value agknkdp.

  (6) At the time of selecting [Pattern 1], only the first KCS learning value agknk among the knock learning values is learned and used to set the required ignition timing afin, so that the ignition timing due to the difference in the octane number of the fuel Can be compensated by the first KCS learning value agnkk. Moreover, since [Pattern 1] to [Pattern 4] are set, all of the knock learning values can be learned at appropriate timings, and the required ignition timing afin is set appropriately based on these knock learning values. be able to.

  (7) As the distribution ratio tk at the time of selecting [Pattern 3], the largest value is set in the medium load operation region, and the smaller the value is set as the distance from the medium load operation region is. Therefore, the distribution ratio tk can be set in accordance with the relationship between the engine load KL and the amount of change in the steady deviation, and the learning update amount tdl is determined as the deposit learning value adep based on the distribution ratio tk. And the first KCS learning value agknk can be appropriately distributed.

  (8) The distribution ratio tk at the time of selecting [Pattern 3] is set to the largest value in the middle rotation operation region, and to a smaller value as the distance from the middle rotation operation region increases. Therefore, the distribution ratio tk can be set in accordance with the relationship between the engine rotational speed NE and the amount of change in the steady deviation, and the learning update amount tdl is set as the deposit learning value based on the distribution ratio tk. It is possible to appropriately distribute to adep and the first KCS learning value agknk.

The embodiment described above may be modified as follows.
The calculation mode of the learning update amount tdl can be arbitrarily changed. For example, a value obtained by performing a gradual change process on the KCS feedback correction value aks can be used as the learning update amount tdl.

  The distribution ratio tk when selecting [Pattern 3] and the distribution ratio tk when selecting [Pattern 4] may be set differently. Further, the distribution ratio tk may be set based on only one of the engine load KL and the engine speed NE. Furthermore, it is possible to set a constant value as the distribution ratio tk regardless of the engine operating state.

  The learning execution mode of the deposit learning value adep can be arbitrarily changed. For example, an arithmetic map for calculating the deposit learning value adep based on the engine operating state may be set, and all the map values of the map may be collectively updated based on the KCS feedback correction value aks. It is also possible to set an arithmetic expression for calculating the deposit learning value adep based on the engine operating state, and to update the coefficient of the arithmetic expression based on the KCS feedback correction value akcs at that time.

  In the above embodiment, the required ignition timing afin is set using the second KCS learning value agnkdp as it is corresponding to the learning region including the current engine operating state among the plurality of learning regions of the multipoint learning region. did. Instead of this, the following values may be used for setting the required ignition timing afin. That is, first, the second KCS learning value aggnkdp corresponding to each learning region in the multipoint learning region is stored as a value corresponding to the representative operation point of each learning region. Then, the second KCS learning value aggnkdp at each representative operating point in the vicinity of the current engine operating state is calculated, and the second KCS learning value agnkdp corresponding to the current engine operating state is calculated through interpolation processing using each representative operating point. To do. Then, the second KCS learning value agnkddp calculated in this way is used for setting the required ignition timing afin.

The apparatus according to the above-described embodiment can be applied to an apparatus in which the entire ratio learning area is included in the multipoint learning area with the configuration appropriately changed.
The apparatus according to the above-described embodiment can be applied to an apparatus in which learning of the first KCS learning value agknk is not performed with the configuration changed as appropriate.

  DESCRIPTION OF SYMBOLS 10 ... Internal combustion engine, 11 ... Cylinder block, 13 ... Combustion chamber, 14 ... Spark plug, 14a ... Igniter, 15 ... Knock sensor, 16 ... Electronic control unit, 17 ... Crank angle sensor, 19 ... Throttle sensor

Claims (8)

The knock control amount is updated according to the occurrence of knocking, and a value for compensating for a steady deviation between the knock control amount and its reference value is learned as a knock learning value, and the knock control amount and the knock learning are learned. An ignition timing control device that sets a control target value for engine ignition timing based on a value,
The knock learning value includes a first knock learning value and a second knock learning value, a first execution region for learning the first knock learning value, and a second execution region for learning the second knock learning value; At least part of
The first execution region is configured to knock the relationship in the entire first execution region based on a change tendency of the relationship between the engine operation state and the first knock learning value with respect to a predetermined change in the knock control amount. It is an area that is updated collectively based on the control amount.
The second execution area is divided according to the engine operating state, and a plurality of areas in which the second knock learning value is set separately are preset, and the engine operating state at that time is included. The ignition timing control device according to claim 1, wherein the ignition timing control device is a region in which the second knock learning value is updated based on the knock control amount. In the ignition timing control device according to claim 1,
In the engine operation region where the first execution region and the second execution region overlap, the first knock learning value and the second knock learning value are learned together and used for setting the control target value,
In the engine operation region in which the first execution region and the second execution region do not overlap, only the value corresponding to the region including the engine operation state at that time is included in the first knock learning value and the second knock learning value. And an ignition timing control device that is used to set the control target value. In the ignition timing control device according to claim 1 or 2,
In the engine operation region where the first execution region and the second execution region overlap, the distribution ratio to the first knock learning value and the second knock learning value for the update for compensating for the deviation is determined by the engine. An ignition timing control device that is set based on an operating state. In the ignition timing control device according to claim 3,
Ignition characterized in that the distribution ratio is a value that is the largest in the engine operating region where the engine load is medium and the value that is distributed to the first knock learning value decreases as the distance from the region increases. Timing control device. In the ignition timing control device according to claim 3 or 4,
The distribution ratio is such that the ratio distributed to the first knock learning value is a value that becomes the largest in an engine operation region where the engine speed is medium and becomes smaller as the distance from the region increases. Ignition timing control device. In the ignition timing control device according to any one of claims 1 to 5,
The knock learning value further includes a third knock learning value,
The third knock learning value is learned in an engine operation region that does not overlap with any of the first execution region and the second execution region, and is used for setting the control target value in all engine operation regions. An ignition timing control device characterized by comprising: In the ignition timing control device according to claim 6,
The apparatus learns the first knock learning value and the third knock learning value in an engine operation region where only the first execution region of the first execution region and the second execution region overlaps. As a ratio for distributing the update amount for compensating the deviation to the first knock learning value in the learning, a value that becomes the largest in the engine operation region where the engine load is medium and becomes smaller as the distance from the region is larger is set. An ignition timing control device characterized by comprising: In the ignition timing control device according to claim 6 or 7,
The apparatus learns the first knock learning value and the third knock learning value in an engine operation region where only the first execution region of the first execution region and the second execution region overlaps. As a ratio for distributing the update for compensating for the deviation to the first knock learning value during the learning, a value that becomes the largest in the engine operation region where the engine speed is medium and becomes smaller as the engine rotational speed is farther away. An ignition timing control device characterized by being set.

Images