JPH0370844A - Knocking control device of internal combustion engine - Google Patents

Abstract

PURPOSE:To increase the reliability of a knocking control device and effectively control knocking by providing a maximum frequent point subtraction updating means for subtracting a determined value from a maximum frequent point and updating when the maximum value of knock signals detected by a knock signal detecting means is within a determined range lower than the most frequent point. CONSTITUTION:The mechanical vibration of the engine body of an internal combustion engine M1 is detected by a knock signal detecting means M2 to output a knock signal, and the most frequent point of the knock signal maximum value in the knock detecting period is determined by a most frequent point calculating means M3. On the basis of the most frequent point of the knock signal maximum value, knock judging level is determined by a knock judging means M4. The knocking is judged by use of this knock judging level. When the knock signal maximum value is within a determined range over the most frequent point, a determined value is added to the maximum frequent point for updating in a most frequent point addition updating means M5. In the converse case, a determined value is subtracted from the maximum frequent point for updating in a subtraction updating means M6.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention (part 1) relates to a knocking control device for an internal combustion engine that corrects a knocking determination level and effectively detects knocking.

[Prior art] Mechanical vibrations that occur in the engine due to abnormal combustion in the internal combustion engine are detected as electrical signals, and when the amplitude of this electrical signal exceeds the target determination level (knock determination level), knocking is detected. Techniques for determining occurrence are known.

This knock detection is performed based on a subtle knock signal that changes depending on the driving condition, so setting the knock detection level described above is extremely important, and various techniques have been proposed to set a more suitable knock detection level ing. Example (1) Using the maximum value (peak) of the amplitude value of the electric signal in a predetermined crank angle range (knock detection period) after ignition of at least one predetermined cylinder, a plurality of ignition cycles of the maximum value A ``knocking detection method for internal combustion engines'' (special) that detects the occurrence of knocking by calculating the average value of , and comparing the knock judgment level based on the average value with the amplitude value of the electric signal during the knock detection period Publication No. 58-28645).

Also, depending on the driving conditions, it may not be possible to set the knock detection level appropriately using method (1).
In recent years, the following technologies have been developed. That is, (2) As shown in FIG. 13, the maximum value of the amplitude value of the knock signal during the knock detection period is determined a predetermined number of times, the most frequent point A is determined from the distribution of the frequency of occurrence, and the coefficient k is set at the most frequent point. ``An engine knocking control device'' (Japanese Patent Application Laid-Open No. 63-253155) that calculates a knocking determination level AXk by multiplying by

(3) As shown in FIG. 13, from the distribution of the frequency of occurrence of the maximum value of the knock signal, find the frequency average value B, which is the average value of the occurrence frequency, and multiply the frequency average value B by the coefficient K. Calculate the judgment level BXK and set this knock judgment level 8.
"Knocking detection device for internal combustion engine" that detects knocking using XK (Japanese Patent Application No. 115133/1983).

[Problems to be solved by the invention] However, the prior art has the following problems and is still not sufficient: (1) The average value of the amplitude value of the electric signal over multiple ignition cycles and the detected This is a configuration that detects the occurrence of knocking by comparing the magnitude of the amplitude value of the electrical signal. When the large amplitude value of the electrical signal is added, the average value when knocking occurs (the average value when knocking does not occur) Therefore, due to the increase in the average value, the knock detection level also cumulatively increases, and unless an electrical signal with a larger amplitude value is input, it becomes impossible to judge whether knocking has occurred, and knocking detection There was a problem that the accuracy deteriorated (2) In addition, the knock detection level AXk was calculated based on the maximum frequency point A of the maximum amplitude value of the knock signal, and the knock detection level AXk was used to detect knocking. Compared to the above example using the average value, the knock judgment level AXk does not vary greatly, so there is an advantage that the accuracy does not decrease when knocks occur frequently. However, the method using the most frequent point A 1.t Since it is necessary to detect the knock signal over a predetermined period of time and perform statistical processing, there is a problem that the amount of data to be stored is large, placing a heavy burden on the RAM, and also requiring a long calculation time (3). 1.t with a configuration in which the frequency average value B of the maximum value of the amplitude value is determined and the knock determination level BXK is set based on the frequency average value B.

Although the fluctuation of the knock detection level BXK is smaller than when using the average value in (1) above, the knock detection level BXK still increases as the occurrence of knocking increases.Especially when knocking occurs frequently, As shown in FIG. 13 (110), the knock determination level AXk is higher than the knock determination level AXk based on the most frequent point A in (2) above,
If detection of knocking becomes difficult, the present invention aims to provide a knocking control device for an internal combustion engine that obtains the most frequent point with a simple configuration and realizes accurate detection of knocking based on the most frequent point.

[Means for solving the problem] The present invention, which has been made to achieve the above object, detects mechanical vibrations of the engine body of an internal combustion engine M as exemplified in FIG. Knock signal detection means M2 that outputs a signal; Most frequency point calculation means M3 that calculates the most frequent point of the maximum value in the knock detection period of the knock signal outputted by the knock signal detection means M2; The most frequent point calculation means M3 Knock determination means M that determines a knock determination level based on the most frequent points determined by and uses the knock determination level to determine the occurrence of knocking.
4. In the knocking control device for an internal combustion engine, the knocking control device comprises: 1. When the maximum value of the knock signal detected by the knock signal detection means M2 is within a predetermined range exceeding the most frequent point; t, most frequency point addition and updating means M5 for updating the maximum frequency point by adding a predetermined value; and a maximum value of the knock signal detected by the knock signal detection means M2 falling within a predetermined range below the most frequency point. In some cases, the gist of the present invention is to provide a knocking control device for an internal combustion engine, comprising: a most frequently occurring point subtracting and updating means M6 that updates the most frequently occurring point by subtracting a predetermined value from the most frequently occurring point.

[Operation] The knocking control device for an internal combustion engine of the present invention (as illustrated in FIG. 1, internal combustion engine M) detects mechanical vibrations of the engine body by the knock signal detection means M2 and outputs a knock signal, Find the most frequent point of the maximum value of the knock signal in the knock detection period by the most frequent point calculation means M3,
Furthermore, the knock determination means M4 determines the most frequent point of the maximum value of the knock signal (based on this, determines the knock determination level,
The occurrence of knocking is determined using this knock determination level. When the maximum value of the knock signal is within a predetermined range exceeding the most frequent point, the most frequent point addition/updating means M5 updates the maximum frequency point by adding a predetermined value. On the other hand, when the maximum value of the knock signal is within a predetermined range below the most frequent point, the most frequent point subtracting and updating means M6 subtracts a predetermined value from the most frequent point and updates it.

In other words, according to the above-mentioned configuration, a suitable maximum frequency point is quickly determined by increasing or decreasing the maximum frequency point according to the maximum value of the knock signal detected during the knock detection period.

Next, the principle of the present invention will be explained in detail based on FIG.

Here, let us consider the case where the distribution of the maximum value of knock signals detected over multiple knock detection periods (two) statistically becomes the graph shown in Fig. 2. In the figure, ■, ■,
■ is the most frequent point AI of the maximum value of the knock signal obtained up to the previous time, A2. A3, and the diagonally shaded areas A1RO, A'guchi', and A'1guchi that extend to the left and right of these indicate the areas where each of the most frequent points A1 to A3 are updated. Furthermore, areas A and B.

A6'' is an area for decreasing correction, and area B2 and B'6'' is an area for increasing correction.

Here, for example, if the previous most frequent point A1 is ■,
In the next measurement, when it is determined that the maximum value of the knock signal is in area A, the most frequent point A1 is corrected to decrease so that the upper When it is determined that it is in the mouth, the most frequent point A1 is corrected by increasing so that (upper part 2) moves to the right side of the graph.

In other words, the maximum value of the knock signal (1) is statistically more likely to appear near the maximum frequency point A in this graph, so the maximum value of the measured knock signal is more likely to be at the area entrance than area A; If the above update process is repeated, the most frequent point AI (the most frequent point A in the graph above) will get closer. This is also true for the other most frequent points A2 and A3, ■ and ■, and each time the update is repeated will approach the most frequent point A on the graph.

In addition, the probability that the maximum value of the knock signal occurs is the frequency average value B
is the same on both sides of the graph, so unless the area to be updated is limited, the calculated most frequent points A1-A3 will not converge to the most frequent point A of the graph, but will converge to the frequency average value B. Therefore, an update area of a predetermined width is set on both sides of the most frequent point A.

In this way, the maximum value of the knock signal and the most frequent point A1-
A3, and by updating the most frequent points A1-A3 one after another based on the latest knock signal, the desired most frequent point A1, which is close to the most frequent point A on the graph, can be quickly found.
−A3 can be obtained. Therefore, this most frequent point A
A knock determination level is obtained by multiplying 1-A3 by a coefficient k, and it becomes possible to accurately detect knocking based on this knock determination level.

The technical problems of the present invention are solved by each component of the present invention acting as described above.

[Example] Next, a preferred example of the present invention will be described in detail based on the drawings. FIG. 3 shows a system configuration of an engine knocking control device according to a first embodiment of the present invention.

As shown in the figure (mi engine knocking control device 1)
It consists of a four-cylinder engine 2 and an electronic control unit (hereinafter simply referred to as ECU) 3 that controls the engine.

Engine 2 (first cylinder (#1) 11, second cylinder (#1)
2) 12, 3rd cylinder (#3) 13, 4th cylinder (#4) 1
4, each cylinder 11. 12. 13゜141 = (Spark plugs 15, 16, 17, 18 are arranged. On these spark plugs 15, 16゜17.18 (, t, igniter with ignition coil] 9
The high voltage necessary for ignition generated in is distributed and supplied via a distributor 20 equipped with a camshaft that interlocks with a crankshaft (not shown).

The engine knock control device 1 includes a resonant knock sensor 3L that is installed in the cylinder block of the engine 2 and outputs mechanical vibration as an electrical knock signal, and a resonant knock sensor 3L that is built in the distributor 20 and installed on the camshaft of the distributor 20. The cylinder discrimination sensor 32 generates a cylinder discrimination signal (G signal) every 174 rotations of the camshaft of the distributor 20, that is, every 180 ['] of the crank angle, and every 1/24 rotation of the camshaft of the distributor 20.
A rotation angle sensor 33 that also serves as a rotation speed sensor that generates a rotation angle signal (Ne signal) every 30 ['] of crank angle, an intake pipe pressure sensor 34 that measures the intake pipe pressure inside the intake manifold, and a cooling water for the engine 2. A water temperature sensor 35 is provided to measure the engine temperature from the temperature. Detection signals from each of these sensors 1i are input to EcU3. ECU3 controls engine 2, MPU3a, and ROM3.
b, RAM3 c, Backup RAM 3
d. Constructed as a logic operation circuit centering around a timer 3e. It is connected to input/output sections 3g and 3h via a common bus 3f to perform input/output with the outside.

The knock signal output by the knock sensor 3 has an impedance conversion effect and has a frequency band specific to knocking (
7-8 [K day Z]) as a pass band Peak hold circuit 3j, MP that performs a holding operation of the maximum amplitude of the knock signal that has passed through the band pass filter circuit 3I according to the control signal of the MPU 3a
According to the control signal of U3a, the output of peak hold circuit 3 is A/D converted, and the A/D conversion end interrupt signal is sent to M.
A/D converter 3k, input/output port 3 outputting to PU3a
The signal is input to the MPU 3a via g. Cylinder discrimination sensor 3
2 detected cylinder discrimination signal (G signal) 1 company Buffer 3
4) The rotation angle signal (Ne signal) detected by the rotation angle sensor 33 is inputted to the interrupt request signal forming circuit 3n via the buffer 3p, the interrupt request signal forming circuit 3n, and the speed signal forming circuit 3q, respectively. The signal is input from the output port 3h to the MPU 3a as an interrupt signal and a rotational speed signal.

Furthermore, the detection signal of the intake pipe pressure sensor 34 is
Detection signals from the water temperature sensor 35 are input to the buffer 3s, and are input from the input/output port 3h to the MPU 3a via the multiplexer 3t and A/D converter 3u, which operate according to the control signal from the MPLJ 3a. On the other hand, MPtJ
3a+A control signal is output to the drive circuit 3v via the input/output port 3g to drive the igniter 9 and control the ignition timing.

Next, FIG. 4 shows the knocking detection start time calculation process executed by the ECU 3, FIG. 5 1n shows the knocking detection end time calculation process, FIG. 6 shows the A/D conversion start process, and FIG. 7 shows the knocking detection process. 1: will be explained based on each.

First, the knocking detection start time calculation process will be explained based on the flowchart shown in FIG. This knocking detection start time calculation process is executed in response to an interrupt signal generated at each predetermined specific crank angle (top dead center (TDC) in this embodiment).

First, in step 100, a process of reading various data is performed. In the following step 110, a process of calculating the knocking detection start time t1 is performed.
Knocking detection start time t1 f1 Start crank angle of predetermined knocking detection period (in this embodiment ('L ATDCIO ~ 20 [''CA])
, the detected current crank angle Cθ['CA], and the current clock value TM of the timer 3e. Next, the process proceeds to step 120, where the knocking detection start time t1 calculated in step 110 is set in the register inside the MPU 3a, and then this process ends. From then on, step 1 for each specific crank angle of this process 1
00~? Repeat step 20.

Next (2) The knocking detection end time calculation process will be explained based on the flowchart shown in FIG. Signal (2) is carried out.

First, step 200 is 1, peak hold circuit 3.”
1: A process of outputting a high level (“1”) control signal is performed. Through the process of step 200, the peak hold circuit 3j[;L starts the peak hold operation of the knock signal. In the following step 210, a process of reading various data is performed.Next, the process proceeds to step 220, and a process of calculating a knocking detection end time t2 is performed.Here, the knocking detection end time t2[t,
The crank angle from the start time to the end time of the predetermined knocking detection period (in this embodiment (for example,
0 to 90 ['cA]), detected current crank angle Cθ['CA], current time value TM of timer 3e
Calculated based on. Proceeding to subsequent step 230, the knocking detection end time t2 calculated in step 220 is determined.
After performing the process of setting the value in the register inside the MPU 3a, the present process ends. Thereafter, this process (steps 200 to 230 are repeated every time an interrupt signal is generated).

Next, the A/D conversion start process for 1 m will be explained based on the flowchart of FIG. This A/D conversion start process is executed in response to an interrupt signal generated at time t2 calculated in the knocking detection end time calculation process described above.

First, in step 300, a process of outputting a high level ("1") control signal to the A/D conversion circuit 3k is performed.
k (Friend) Starts A/D conversion of the knock signal. After performing the process of step 300, -1 the process ends.

Thereafter, this process (step 300) is repeated every time an interrupt signal is generated.

Next 1: The most frequent point update process used for knocking detection will be explained based on the flowchart of FIG. In this most frequent point update process (knocking detection process (steps 400 to 440)) and the most frequent point update process (step 445), which is the main part of this embodiment,
to 470) will be explained together. In this process 11, the A/D conversion end interrupt signal outputted to the A/D converter 3 (in this embodiment, 10 [m5e
c] Occurs during elapsed time).

First, in step 400, a process of reading various data is performed.In subsequent step 405, a process of setting the A/D conversion value A/D to the knock sensor output signal amplitude value (knock signal amplitude value) a is performed. then step 4
Proceeding to step 10, processing is performed to output a peak hold end control signal to the peak hold circuit 3. That is, M
P U 3 a Lt.

A low level ("O") control signal is output to the peak hold circuit 3J.

In the following step 415, it is determined whether the knock signal amplitude value a is less than or equal to the knock determination level xA for determining the occurrence of knocking, that is, whether or not knocking has occurred.
If it is determined that this has not occurred, the process proceeds to step 420,
On the other hand, if it is determined that this has occurred, the process proceeds to step 425.

Here, the value is a coefficient (in this example (for example, 1ik=2
), and the value A is the maximum frequency point of the knock signal amplitude value a, and is updated sequentially.

In the above step 420, it is determined whether the count value of the counter n that counts the number of consecutive times of no knocking is greater than or equal to the value 10, and if an affirmative determination is made, the process proceeds to step 430,
On the other hand, if the determination is negative, the process advances to step 435.

In step 435, which is executed when the knocking-free state is less than 10 consecutive ignition cycles (after performing the process of adding 1 to the count value of the upper counter n, step 44
Proceed to step 5.

On the other hand, if the state without knocking is 10 consecutive ignition cycles
In step 430, which is executed when the ignition timing is greater than or equal to 1, the ignition timing advance correction value θ is advanced by the crank angle Y. This advance angle correction value θ is a value used in the well-known ignition timing calculation process that is executed at every predetermined crank angle. Based on this, the basic ignition timing θO is calculated according to a map determined in advance and stored in the ROM 3b, and the basic ignition timing θO is corrected by the advance angle correction value θ to calculate the target ignition timing θ. The target ignition timing θ is determined by the process in step 430.
・(Top) The crank angle is advanced by Y.

In the subsequent step 440, the counter n is reset to the value O, and then the process proceeds to step 445.

Furthermore, in step 425, which is executed when knocking occurs, the ignition timing advance correction value θ is retarded by the crank angle X.

This process retards the target ignition timing θ' (upward by the crank angle X) and suppresses the occurrence of knocking.

Thereafter, the process proceeds to step 445 via step 440.

In the following step 445, it is determined whether the knock signal amplitude value a is less than or equal to the product of the most frequent point A and the coefficient kl, and if an affirmative determination is made, the process proceeds to step 450. kl=1. 2) The knock signal amplitude value a at t is within a predetermined range exceeding the most frequent point A (for example, within the second
This value is used to determine whether or not the knock signal amplitude value a is in the region of 9° 1 mouth" in the figure. If the knock signal amplitude value a exceeds the above range, the most frequent point A is updated. do not have.

In the following step 450, contrary to step 445 above, it is determined whether the knock signal amplitude value a is on the loading dock with the most frequent point A and the coefficient 2, and if an affirmative determination is made, step 455 Proceed to .Here, set the coefficient to 2 (for example, 2=0
．． 8) t;i, the knock signal amplitude value a is within a predetermined range below the most frequent point A (for example,
The most frequent point A is not updated if it is below this range.

In other words, the processing in steps 445 and 450 above (the processing for determining whether the knock signal amplitude value a detected above is in a certain update area including the most frequent point A, and only when it is in this range, the following The most frequent point A is updated in step 455 and the like.

Then, in step 455 (Dai), in order to update the most frequent point A, it is determined whether the knock signal amplitude value a exceeds the most frequent point A. If an affirmative determination is made, the process proceeds to step 460; If determined, the process advances to step 465.

In the step 460 described above, after performing the process of increasing the correction by adding the value 1 to the most frequent point A, the process proceeds to step 470. On the other hand, in the step 465, the value 1 is subtracted from the most frequent point A to decrease the After performing the correction process, step 4
Proceed to 70.

In step 470, various updated data are stored in RAM 3c or backup RAM.
After performing the process of storing data in 3d, this process ends on -1.

As explained above, according to the first embodiment, the knock signal amplitude value a obtained by A/D converting the peak hold value of the output signal of the knock sensor 31 is within a predetermined range exceeding the most frequent point A. When the knock signal amplitude value a is within a predetermined range below the most frequent point A, the value 1 is added to the most frequent point A to correct the increase.
The decrease is corrected by subtracting . As a result, the most frequent point A is successively corrected, so it is possible to quickly and suitably set the optimum most frequent point A according to the driving condition.
Using this most frequent point A, it becomes possible to accurately detect knocking. Furthermore, since there is no need to store and statistically process a large amount of data in order to calculate the most frequent points, the area used for storing data such as the RAM 3c can be reduced, and the statistics of the MPU 3a can be reduced. This has the advantage of reducing the computational burden of processing.

In addition, since the most frequent point A is used to calculate the knock detection level, even if knocking occurs frequently, the amount of variation in the knock detection level is smaller than when the frequency average value B is used, resulting in stable knocking. Occurrence can be determined.

Next, a second embodiment of the present invention will be described in detail based on the drawings. Main differences between this second embodiment and the first embodiment (friends)
This embodiment uses not only the most frequent point A but also the area average value (frequency average value) B of the frequency distribution, and depending on the increase or decrease in the occurrence of knocking, the knock judgment level is set based on the most frequent point A or the frequency average. The purpose is to switch to one based on value B and use it. Since the other device configurations are the same as those of the first embodiment, the same parts are denoted by the same reference numerals and the explanation will be omitted.

The most frequent point update process executed in the second embodiment is
This will be explained based on the flowchart shown in the figure.

This process (like the first embodiment) consists of a knocking detection process (steps 500 to 520) and a most frequent point updating process (steps 522 to 544).

First, in step 500, a process of reading various data is performed. In the subsequent step 502, processing is performed to set the A/D conversion value A/D to the knock signal amplitude value a. Next, the process proceeds to step 504, where the peak hold circuit 3
Processing is performed to output a peak hold end control signal to j.

In the following step 506, it is determined whether the knock determination bell KXB based on the frequency average value B exceeds the knock determination level 'XBXM based on the most frequent frequency point A,
If the determination is affirmative, the process proceeds to step 508, while if the determination is negative, the process proceeds to step 5]O. Here, the coefficient (Y) is a coefficient that converts the frequency average value B into a knock judgment level, and the coefficient '' is a coefficient set to a slightly larger value than the coefficient of the first embodiment, and the coefficient M and is the value A/B calculated using the most frequent point A and the frequency average value B set last time.

The determination in step 506 (1) As shown in FIG. This process compares the knock detection level with 'XBXM and selects the knock detection level that is most suitable for detecting knocking.
Proceeding to step 508, the knock judgment level is set based on the most frequent point A, which has little fluctuation even when knocking occurs frequently.
Use XBXM. On the other hand, when the occurrence of knocking is small (above), the process proceeds to step 510 and uses the knock determination level KXB based on the frequency average value B, which has excellent transient response.

In the above step 508, it is determined whether the knock signal amplitude value a is less than or equal to the knock determination level 'XBXM based on the most frequent point A to determine whether knocking has occurred, and it is determined that knocking has not occurred. and step 512
On the other hand, if it is determined that knocking has occurred, the process advances to step 514.

In the above step 512, it is determined whether the count value of the counter n that counts the number of consecutive times in which there is no knocking is greater than or equal to the value 10, and if an affirmative determination is made, the process proceeds to step 516,
On the other hand, if the determination is negative, the process advances to step 518.

Then, the state without knocking is continuous ignition cycle 1.
In step 518, which is executed when the number of times is less than 0, the process of adding 1 to the count value of the counter n is performed, and then the process proceeds to step 522.

On the other hand, in step 516, which is executed when there are 10 or more consecutive ignition cycles without knocking, the ignition timing advance correction value θ is advanced by the crank angle Y. Therefore, in step 516 As a result of the process, the target ignition timing θ° (y) is advanced by the crank angle Y. In the following step 520, after the counter n is reset to the value O, the process proceeds to step 522.

Furthermore, in step 514, which is executed when knocking occurs, processing is performed to retard the ignition timing advance correction value θ by the crank angle The occurrence of knocking is suppressed only when the timing is retarded.

Thereafter, the process proceeds to step 522 via step 520.

Next, in step 522, it is determined whether or not the frequency average value B is greater than or equal to the knock signal amplitude value a.If the determination is affirmative, the process proceeds to step 524, while if the determination is negative, the process proceeds to step 526.

If the frequency average value B is greater than or equal to the knock signal amplitude value a, it is determined in step 524 whether or not the frequency average value B is equal to the knock signal amplitude value a, and if the frequency average value B is equal, the process directly proceeds to step 530. If they are not equal, the value 1 is subtracted from the frequency average value B in step 528 to correct the decrease, and the process proceeds to step 530.

On the other hand, if the frequency average value B is not equal to or greater than the knock signal amplitude value a, one company adds the value 1 to the frequency average value B in step 526 to perform an increase correction, and the process proceeds to step 530.

In other words, in the above steps 522 and 524, it is determined that the frequency average value B exceeds the currently measured knock signal amplitude value a (1) The frequency average value B is biased towards the larger knock signal amplitude value a Therefore, in step 528, the frequency average value B is corrected to decrease by taking this current data into account.
If the frequency average value B is lower than the knock signal amplitude value a (this indicates that the frequency average value B is biased toward the smaller side of the knock signal amplitude value a), an increase is corrected in step 526. be.

In the following step 530, the above coefficient M is added to the frequency average value B.
Processing is performed to set the value multiplied by the value as the provisional most frequent point. In other words, by multiplying the value A/B of this coefficient M by the current frequency average value B that has been corrected to decrease or increase,
Set a temporary most frequent point.

In the following step 532, it is determined whether or not the knock signal amplitude value a is multiplied by a coefficient of 3 (for example, 3=1°2), and an affirmative determination is made. Then, in step 534, the knock signal amplitude value a is the same as the most frequent point A and the coefficient is 4 (for example, 4:o, 8). It is determined whether or not the above is the same, and if it is determined in the affirmative, step 5
Proceed to step 36.

In other words, the processes in steps 532 and 534 (similar to the first embodiment) are processes for determining whether or not the knock signal amplitude value a is within the range for updating the most frequent point A.

In step 536, it is determined whether or not the knock signal amplitude value a exceeds the provisional most frequent frequency point, and if the determination is affirmative, the process proceeds to step 538, while if the determination is negative, the process proceeds to step 540.

Then, in step 538, which is executed when the knock signal amplitude value a is on the larger side than the tentative highest frequency point
After performing an increase correction process by adding a value of 0.01 to the coefficient M, the process proceeds to step 542.

On the other hand, in step 540 (upper
After performing a process of subtracting the value 0.01 from the coefficient M to perform a reduction correction, the process proceeds to step 542.

In other words, the coefficient M for updating the most frequent point A is corrected to decrease or increase in accordance with the determination based on the latest knock signal amplitude value a in step 536.

In the following step 542, the frequency average value B is multiplied by the corrected coefficient M, and that value is set as the most frequent point A. In the following step 544, the value and various data are stored in the RAM 3.
c, or after performing the process of storing it in the backup RAM 3d, the sample process ends.

As explained above, according to the second embodiment, in addition to the effects of the first embodiment described above, the following unique effects are achieved.

In the second embodiment, the most frequent point A is determined using the coefficient M, which is the ratio of the most frequent point A and the frequency average value B.
It can also be used when the rotation of the engine 2 increases (:, the knock signal amplitude value a increases and the entire frequency distribution changes to the side with a larger output. In other words, the maximum of all the knock signal amplitude values a) Since the frequency average value B that is updated using the value responds in accordance with the frequency distribution that changes depending on the driving state, the most frequent point A also changes together, making it possible to suitably follow transients.

Furthermore, when the occurrence of knocking increases, the knock judgment level based on the most frequent point A, which is stable against changes, is used; on the other hand, when the occurrence of knocking decreases (transient response Since the knock determination level based on the frequency average value B is used, the knock determination level can be suitably switched depending on the occurrence status of knocking, and the occurrence of knocking can be accurately detected.

Note that as the process of updating the most frequent point A in the first embodiment (steps 445 to 470), the process of updating the most frequent point A in the second embodiment (steps 522 to 5)
44) may be adopted, or conversely, the process of updating the most frequent point A in the second embodiment (step 522) may be adopted.
~544), the process of updating the most frequent point A in the first embodiment (steps 445~470) may be adopted. In other words, the process of updating the most frequent point A in the first embodiment has the advantage of easy calculation, and the process of updating the most frequently occurring point A in the second embodiment has the advantage of excellent responsiveness during transient times. There are two types, so you can choose the features of either one.

Next, a third example and a fourth example will be described. The third embodiment and the fourth embodiment are different from the first embodiment or the second embodiment.
Significant difference from the embodiment 1. The most frequent point A is divided into an area for increasing correction and an area for decreasing correction, and each area is divided into an area for large correction and an area for small correction.

First, the most frequent point update process of the third embodiment will be explained based on the flowchart of FIG. 9.

In step 600 (above), processing to read various data is performed.In subsequent step 605, the t&A/D conversion value A/
A process is performed to set D to the knock signal amplitude value a.

Next, the process proceeds to step 610, where a process of outputting a peak hold end control signal to the peak hold circuit 3J is performed.

In the following step 615, it is determined whether or not the most frequent point A obtained up to the previous time is greater than or equal to the knock signal amplitude value a, and if an affirmative determination is made, the process proceeds to step 620, whereas if a negative determination is made, the process proceeds to step 620. Proceed to 625. Step 620
[飄] It is determined whether the most frequent point A is equal to the knock signal amplitude value a, and if they are equal, the process proceeds to step 655;
On the other hand, if they are not equal, the process proceeds to step 630.

In this step 630, it is determined whether the knock signal amplitude value a is less than the product of the most frequent point A and a coefficient of 5 (for example, 5=0.8), and if an affirmative determination is made, step 63
If the determination is negative, the process advances to step 640. In this step 635, after performing the process of subtracting the value 1 from the friend most frequent point A to correct the decrease, step 655
Proceed to. On the other hand, in step 640, the value 2 is subtracted from the most frequent point A to correct the decrease, and then the process proceeds to step 655.

Also, in step 625 above (friend knock signal amplitude value a
It is determined whether or not exceeds the product of the most frequent point A and the coefficient 6 (for example, 6=1°2), and if the affirmative determination is made, the process proceeds to step 645, while if the negative determination is made, the process proceeds to step 650.
Proceed to. In this step 645, the value 1 is added to the most frequent point A to correct the increase, and then step 6
Proceed to step 55.

On the other hand, in step 650, a value 2 is added to the most frequent point A to perform an increase correction process, and then the process proceeds to step 655.

Here, the processing of steps 615 to 650 (the 10th
As shown in the figure (:, first, it is determined whether the knock signal amplitude value a is in the left or right region of the most frequent frequency point A obtained up to the previous time, and it is determined whether to perform a decreasing correction or an increasing correction. Next, it is determined whether or not the knock signal amplitude value a is within a predetermined range shown by diagonal lines in FIG.
Within that range, the update amount of the most frequent point A is increased to 2,
In the case of an outer area that protrudes from that area (see E and F in the figure), the update amount is set to a small value of 1.

In the following step 655, it is determined whether the knock signal amplitude value a is equal to or lower than the knock determination level kXA for determining the occurrence of knocking, and if an affirmative determination is made, step 660
On the other hand, if the determination is negative, the process advances to step 665.

In step 660, which is executed when no knocking occurs, it is determined whether the count value of a counter n that counts the number of consecutive times in which no knocking occurs is greater than or equal to the value 10,
If the determination is affirmative, the process proceeds to step 670, while if the determination is negative, the process proceeds to step 675.

In step 675, which is executed when the knock-free state is less than 10 consecutive ignition cycles (after performing the process of adding the value 1 to the count value of counter n, step 68
Proceed to step 5.

On the other hand, in step 670, which is executed when there are 10 or more consecutive ignition cycles without knocking, a process is performed in which the ignition timing advance correction value θ is advanced by the crank angle Y.

In the following step 680, the counter n is reset to the value O, and then the process proceeds to step 685. Also, step 66 is executed when knocking occurs.
In step 5, a process is performed to retard the ignition timing advance angle correction value θ by the crank angle X in order to suppress knocking. Thereafter, the process proceeds to step 685 via step 680.

In this step 685, the updated data can be stored in RAM 3c or backup RAM.
After performing the process of storing data in 3d, this process is immediately terminated.

As explained above, according to the third embodiment, the area where the most frequent point A is corrected to increase and the area where the most frequent point A is corrected to decrease are each divided into two areas, and the amount of update in the area closer to the most frequent point A is divided into two areas. is larger than the amount of updates in the distant area (twice as much)
It is set. In this way, by setting different update amounts, even if the knock signal amplitude value a fluctuates greatly during a transient period, the most frequent point A can be quickly updated in accordance with changes in the operating condition, thereby increasing the knock judgment level. The occurrence of knocking can be suitably detected by changing.

In other words, in this embodiment, since the update is performed not only in a part of the area but in the entire area, the most frequent point A has the advantage of being quickly updated and has excellent followability. Since the update amount of is increased, there is an advantage that the most frequent point A is prevented from fluctuating excessively and convergence is excellent.

Furthermore, when using the coefficient M mentioned above, if the operating condition suddenly changes, the statistical distribution will also change significantly, so in some cases the guard may be tightened and the most frequent point A may be reset from the beginning again.
However, in this embodiment, the most frequent point A is quickly changed over the entire area, which reduces the possibility of guarding and simplifies the calculation. be.

[Next]: A fourth embodiment will be described based on FIGS. 11 and 12.

This fourth embodiment differs greatly from the third embodiment described above.
The purpose is to change the update amount of the most frequent point A in accordance with changes in the engine speed. A guard is applied to prevent this amount of change from becoming too large.

First, the guard value update process in this embodiment will be explained based on the flowchart of FIG. 11. This process is executed by an interrupt every 200 m5ec.

In step 700, subtract the current engine speed Nnew from the previous engine speed No1d determined based on the signal from the rotational speed sensor 33, and the difference is 1150.
The absolute value of is set as the correction value d. The reason for dividing by 50 here is that simply using the difference in engine speed would result in too large an update amount, making it impossible to obtain the most frequent point A in a suitable manner.

Next 1: In step 710, the current engine speed Nnew is set to the previous engine speed No1d.

In the following step 720, in order to update the guard value C that checks the correction value dt, it is determined whether the correction value d is greater than or equal to the previously set guard value C, that is, whether the change in engine speed is large. , if a positive determination is made, step 730
On the other hand, if the determination is negative, the process advances to step 740. Then, in step 730, it is assumed that the change in the engine speed is increasing, and the correction value d is set to the guard value C. On the other hand, in step 740, it is assumed that the change in the engine speed is converging. and set the guard value C to 1/2 of that
) Second correction.

In other words, if the amount of change in engine speed becomes large, the guard value C is also increased, but if the amount of change is small, the guard value C is decreased. In the example (the change in engine speed is 5
Orpm/ 200 m5ec or less becomes d],
Since d=O due to digit loss, C=0. Therefore, the calculated most frequent point A when there is no rotational change converges to the statistical most frequent point A.

Next, the most frequent point update process in this embodiment will be explained based on the flowchart of FIG. 12.

In step 800 (above), processing to read various data is performed.In subsequent step 805, +tA/D conversion value A/
A process is performed to set D to the knock signal amplitude value a.

Next, the process proceeds to step 810, where processing is performed to output a peak hold end control signal to the peak hold circuit 3).

In the following step 815, processing is performed to set the absolute value of 1/4 of the difference between the most frequent point A and the knock signal amplitude value a as the correction value d. This is to prevent the most frequent point A from being updated unnecessarily when the signal amplitude value a is small.

In step 820, it is determined whether the correction value d is less than or equal to the guard value C, and if an affirmative determination is made here, step 830
On the other hand, if a negative determination is made, the process proceeds to step 825, where the value of the guard value C is set as the correction value d. In other words,
If it is determined in step 820 that the correction value d is not less than the guard value C, the correction value d is set using the guard value C obtained in the process shown in FIG. 11 described above.

In the following step 830, it is determined whether or not the most frequent point A obtained up to the previous time is greater than or equal to the knock signal amplitude value a, and if an affirmative determination is made, the process proceeds to step 835, whereas if a negative determination is made, the process proceeds to step 835. Proceed to 840. Step 835
(Y) It is determined whether the most frequent point A is equal to the knock signal amplitude value a, and if they are equal, the process proceeds to step 870;
On the other hand, if they are not equal, the process advances to step 845.

In this step 845, it is determined whether the knock signal amplitude value a of one company is less than the product of the most frequent point A and a coefficient of 7 (for example, 7=0.8), and if an affirmative determination is made, step 855 is performed.
If the determination is negative, the process advances to step 850.

Then, in step 850, after performing a process of subtracting the value 1 from the upper most frequent point A to correct the decrease, the process proceeds to step 855.

In this step 855 (from the most frequent point A to the correction value d
The value obtained by subtracting is set as the most frequent point A. That is, in steps 845 to 855, the knock signal amplitude value a
In a predetermined range far from the most frequent point A (for example, the area E in Fig. 10), the most frequent point A is only reduced by the correction value d, but the knock signal amplitude value a is far from the most frequent point A. In a predetermined range close to (for example, the area C in Fig. 10) (
, t, the value from the most frequent point A] and the correction value d.

Also, in step 840 above, the knock signal amplitude value a
It is determined whether or not exceeds the product of the most frequent point A and the coefficient kg (for example, 8=1°2). If the determination is affirmative, the process proceeds to step 865, while if the determination is negative, the process proceeds to step 850. In this step 860, the most frequent point A has a value of 1.
After performing the process of adding and correcting the increase, step 86
Proceed to step 5. In step 865, the value obtained by subtracting the correction value d from the most frequent point A is set as the most frequent point A. That is, in steps 840, 860, and 865, the knock signal amplitude is In a predetermined range where the value a is far from the most frequent point A (for example, the area shown in FIG. In a predetermined range near point A (for example, area 2 in Fig. 10),
A large incremental correction is performed by adding the value 1 and the correction value d to the most frequent point A.

In the following step 870, the data is stored in the RAM 30, etc., and then the process is temporarily terminated.

As explained above, according to the fourth embodiment, in addition to the effect of excellent convergence to the most frequent point as in the third embodiment, the most frequent point A can be adjusted according to the engine speed. Since the correction value d is changed, the followability during transition is excellent. Further 1:.

Since the guard value C is changed and set according to the engine speed, it is possible to set a suitable guard value C during a transient period. Thereby, the most frequent point A can be suitably adjusted, and the large knock generation formula can prevent the occurrence of false retardation and the like.

Although several embodiments of the present invention have been described above, the present invention is not limited to these embodiments in any way, and can be implemented in various ways without departing from the gist of the present invention. Of course.

Effects of the Invention As described in detail above, when the maximum value of the knock signal is within a predetermined range exceeding the most frequent point (Table 1), the knocking control device for an internal combustion engine of the present invention Moreover, when the maximum value of the knock signal is within a predetermined range below the most frequent point, the configuration is such that the predetermined value is subtracted from the most frequent point and updated.Therefore, the calculation for updating the most frequent point is easy. This has the excellent effect of reducing the burden placed on the processing device.In other words, it does not necessarily require large-scale storage capacity or high-speed processing power, so it can be realized with a relatively simple device configuration, and is highly versatile and easy to install in vehicles. Can improve sex.

In addition, the difference in the most frequent points between when knocking occurs and when knocking does not occur is relatively small, and the knock judgment level for detecting knocking does not vary greatly, increasing the reliability of the knocking control device and Execution of effective one-knocking suppression control has the effect of increasing the durability of the internal combustion engine.

[Brief explanation of drawings]

Fig. 1 shows the basic configuration conceptually illustrating the content of the present invention, Fig. 2 is a graph explaining the invention in detail, and Fig. 3 shows the system configuration of the first embodiment. 6 and 7 are flow charts showing the same control, and FIG.
FIG. 9 is a flowchart showing the control of the second embodiment, FIG. 9 is a flowchart showing the control of the third embodiment, and FIG. 10 is a graph showing the relationship between the knock signal amplitude value and the frequency of occurrence.
FIG. 11 is a flowchart showing the guard value update process of the fourth embodiment, FIG. 12 is a flowchart showing the most frequent point update process, and FIG. 13 is a graph explaining the prior art. 3 4 5 6 a 1 Maximum frequency point calculation means Knock determination means Maximum frequency point addition and updating means Maximum frequency point subtraction updating means Engine knocking control device Engine electronic control unit (ECU) PU knock sensor

Claims (1)

[Scope of Claims] 1. Knock signal detection means for detecting mechanical vibrations of the engine body of an internal combustion engine and outputting a knock signal, and a maximum value of the knock signal outputted by the knock signal detection means during a knock detection period. a most frequent point calculation means for calculating a frequency point; a knock determination method that calculates a knock determination level based on the most frequent point calculated by the most frequent point calculation means, and uses the knock determination level to determine whether knocking has occurred; In the knocking control device for an internal combustion engine, if the maximum value of the knock signal detected by the knock signal detection means is within a predetermined range exceeding the most frequent point, If the maximum value of the knock signal detected by the most frequent point addition update means that adds and updates the values and the knock signal detection means is within a predetermined range below the most frequent point, the maximum frequency 1. A knocking control device for an internal combustion engine, comprising: maximum frequency point subtraction updating means for subtracting and updating a predetermined value from a point.