JP2889651B2 - Misfire detection device for internal combustion engine - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a misfire detection device that detects a misfire of an internal combustion engine from a minute change in the angular speed of a crankshaft.

2. Description of the Related Art When a misfire occurs in an internal combustion engine, unburned air-fuel mixture is discharged as it is, which not only causes an increase in harmful components in the exhaust gas, but also causes a decrease in output and a decrease in engine stability. Invite.
Therefore, it is easy to determine whether a misfire has occurred in recent years.
For example, there is a demand for a misfire detection device that detects a fire in a normal operation state.

As one of the methods for detecting the misfire, a method of detecting a minute change in the angular velocity of the crankshaft by using, for example, an electromagnetic pickup or the like has been considered. For example, JP-A-57-188748
According to the publication, a rotating plate having a projection piece is fixed to a crankshaft at every 180 ° of crank angle, and proximity switches are provided at two places with phases different by 90 °. There is shown an apparatus for detecting the presence or absence of an error.

However, in the above-described conventional configuration, it is very difficult to reliably detect a small speed difference due to a misfire because it is trying to detect the time required for the protrusion to pass through the front surface of the proximity switch. In particular, in a state where the engine is operated at a certain rotational speed other than at the time of idling, the detection becomes substantially impossible.

In the past, the details of misfires, that is, whether misfires occurred frequently in specific cylinders, whether misfires occurred in multiple cylinders, or very rarely caused misfires Identification is not considered.

Further, since the angular velocity fluctuation is relative, if the angular velocity changes due to other factors such as friction, it is greatly affected by the change, and sufficient detection accuracy cannot be obtained.

As shown in FIG. 1, a misfire detection device for an internal combustion engine according to the present invention includes: a reference position detection unit 1 that outputs a reference position signal at a reference crank angle position of each cylinder; A sensor 2 for outputting a predetermined signal for each angle, and a means 3 for measuring a time required for detecting a predetermined number of signals output from the sensor from a predetermined phase based on the reference position signal
Means 4 for calculating the degree of roughness of each cylinder by dividing the deviation of the required time sequentially measured by the average value of the required time. The average value of the absolute value of the degree of roughness of all cylinders is a predetermined reference value. Means 5 for counting the number of times exceeded to obtain a first parameter, means 6 for obtaining a second parameter based on the maximum value of the absolute value of the roughness within a predetermined period, and means for calculating the roughness of each cylinder. Means 7 for adding and subtracting the number of positive and negative times for a predetermined period, and obtaining a third parameter as a sum of the absolute values; based on these first to third parameters, a frequent misfire occurrence state of one cylinder and a frequent misfire Means 8 for distinguishing between a misfire occurrence state and a low frequency misfire occurrence state.

As shown in FIG. 2, means 9 for obtaining a roughness degree learning value as a moving average for each cylinder every time the roughness degree is calculated.
And means 10 for correcting the degree of roughness calculated sequentially using the roughness part learning value.

The required time required to detect the predetermined number of signals is proportional to the angular velocity of the crankshaft.

The roughness degree is defined as the deviation of each required time / the average value of the required time, so if there is no misfire and a constant angular velocity is obtained during the explosion stroke of each cylinder, the roughness degree becomes 0,
If there is a misfire, the roughness detected at the corresponding phase has a negative value. Then, the roughness detected at the phase corresponding to the other cylinder has a relatively positive value.

Therefore, the first parameter obtained by counting the number of times that the average value of the absolute values of the roughness degrees exceeds the predetermined reference value indicates the degree of rotation fluctuation of the entire engine.

In addition, the second parameter indicates whether or not a large variation to some extent, that is, a large variation due to misfire occurs even at a low frequency.

The third parameter is obtained by adding or subtracting the positive / negative number of roughness degrees and calculating the sum of the absolute values. Therefore, if misfiring occurs more randomly in multiple cylinders than in the case of misfiring in one cylinder, the third parameter is positive or negative. Shows a relatively small value due to the cancellation of

Therefore, a frequent misfire occurrence state of one cylinder, a frequent misfire occurrence state of a plurality of cylinders, and a low-frequency misfire occurrence state can be identified by a combination of the first to third parameters.

In the configuration shown in FIG. 2 including the learning value calculation means 9 and the correction means 10, when a ring gear is used as a sensor, variations in required times caused by processing accuracy of the ring gear are corrected.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 3 is an explanatory view showing a mechanical configuration of one embodiment of the present invention, in which 11 is an internal combustion engine, 12 is an intake passage, and 13 is an exhaust passage. In this embodiment, a 4-cycle in-line 6-cylinder internal combustion engine will be described as an example.

In the intake passage 12, a fuel injection valve 14 for supplying fuel toward each intake port is provided for each cylinder, and a throttle valve 15 is interposed.The throttle valve 15 is provided upstream of the throttle valve 15. ,
For example, a hot wire type air flow meter for detecting the intake air amount Q
Sixteen are arranged. The throttle valve 15 is provided with a throttle valve opening sensor 17 for detecting the opening TVO.

 Reference numeral 18 denotes a vehicle speed sensor that detects a vehicle speed VSP.

Reference numeral 19 denotes a crank angle sensor provided at the end of the camshaft, inside the distributor, or the like to detect the engine speed Ne and the crank angle position. The crank angle sensor 19 includes a pulse signal (REF signal) for detecting a reference position of each cylinder, for example, a top dead center position, and a unit crank angle (for example, 2 ° CA) for detecting a rotation angle from the reference position. ) And a pulse signal (POS signal). The REF signal is output at every 120 ° CA in the case of a six-cylinder engine, and is output at a predetermined angle before the top dead center of each cylinder. More specifically, the pulse width of the pulse corresponding to each cylinder is different.
It is possible to detect the compression top dead center position of the cylinder, and thus to discriminate the cylinder of each cylinder (see FIG. 4 (a)). Therefore,
In this embodiment, the REF signal is used as a reference position signal.

On the other hand, a ring gear 21 that can mesh with a starter motor (not shown) is attached to the rear end of the crankshaft 20 together with a flywheel. The teeth of the ring gear 21 are formed at equal intervals, for example, every 6 ° CA. A ring gear sensor 22 including an electromagnetic pickup or the like is provided near the ring gear 21. The ring gear sensor 22 is configured to form an AC current generated by the passage of the teeth of the ring gear 21 into an ON-OFF pulse signal and output the ON / OFF pulse signal, thereby obtaining a pulse train as shown in FIG. Can be When the crank angle sensor 19 is provided for a camshaft or the like, the accuracy of the POS signal is affected by the backlash of the valve operating mechanism or the like.
It is difficult to detect angular velocity fluctuation due to misfire from the signal. On the other hand, since the ring gear sensor 22 detects the rotation speed of the ring gear 21 directly connected to the crankshaft 20, it is possible to reliably detect angular velocity fluctuation.

The control unit 23 to which the detection signals of the above-described sensors are input uses a so-called microcomputer system, and performs the air-fuel ratio control by the fuel injection valve 14, the ignition timing control of an ignition system (not shown), and the like. Performing a misfire determination based on the detection signal of the ring gear sensor 22,
When a predetermined misfire is detected, a failure code corresponding to the predetermined misfire is stored, and alarm means such as an alarm lamp 24 is activated. Further, the failure code can be read by external means.

 Next, the operation of the above embodiment will be described.

First, the detection of the angular velocity for each cylinder will be described with reference to FIG. 4. As shown in FIG.
After detecting the REF signal output every time, the ring gear sensor 22
Is started, and when a predetermined number of teeth, for example, P teeth are detected, the control unit 23
A first trigger as shown in FIG. 4C is output by a preset counter (not shown). Furthermore, from now on
When the number of teeth is detected, a preset counter (not shown) outputs a second trigger as shown in FIG. That is, during this time, the crankshaft 20 rotates by the crank angle corresponding to the number of m teeth. Then, the timer in the control unit 23 measures the required time T of the engine from the first trigger to the second trigger.

The measurement of the required time T is repeatedly performed every time the REF signal is output, that is, every 120 ° CA. And the number of cylinders is N
(N + 1) data is always stored as (N = 6 in this embodiment). More specifically, the data up to T (N + 1) is stored in a sequentially updated manner, with the latest data being T1 and the previous data being T2, and a misfire determination is made based on these, as described later. .

The phase for measuring the required time T should, of course, be set by selecting a period in which the angular velocity fluctuation due to misfire appears most remarkably. In this way, by focusing only on the phase within a certain range, there is no influence even if there is a change in angular velocity due to friction or the like in other periods.

Next, FIG. 5 is a main flowchart showing a specific processing flow in the control unit 23.
First, it is determined whether or not various conditions are within a diagnosis area where misfire can be determined (step 1). Only when the conditions are within the diagnosis area, the process proceeds to step 3 and subsequent steps. If it is out of the diagnostic area, the measurement of the required time T is stopped (step 2), and the TRAVMX described later is used.
Are cleared (step 13). The determination as to whether or not it is within the diagnostic region is basically a determination as to whether or not there is another factor such as a change in the angular velocity of the engine, and various methods can be considered. An example is shown in FIG. That is, the change amount ΔVSP of the vehicle speed VSP is equal to or smaller than the predetermined value ΔVSP1 (step 21), and the change amount ΔT of the throttle valve opening TVO is set.
VO is equal to or less than a predetermined value ΔTVO1 (step 22), the starter switch is OFF (step 23), and the electric load is constant, that is, not immediately after the ON / OFF change of the headlight switch (step 22). 24) that the power steering load is constant, that is, the power steering operation is not being performed (step 25), that the engine speed Ne is substantially constant, that is, that the engine speed Ne is within a predetermined range of Ne1 to Ne2 (step 26);
The load of the engine, for example, that the basic fuel injection amount Tp is substantially constant, that is, within a predetermined range of Tp1 to Tp2 (step 27) is set as a diagnostic condition. (Step 29) If any one of the conditions is not satisfied, it is determined to be out of the diagnostic region (Step 29).
28).

If it is determined in the above processing that the time is within the diagnostic area, the process proceeds to step 3 and the measurement of the required time T described above is started.

Then, the roughness degree RO is sequentially calculated based on the measured required time T, specifically, the data of T1 to T (N + 1) held at that time (step 4).

This roughness degree RO is obtained by calculating the deviation of the required time T from the required time T
Divided by the average of, for example, Is required. This example focuses on T (N / 2) which is the center point of the data of T1 to T (N + 1).
T (N +) one cycle before the same cylinder as before (720 ° CA before)
By using 1), it is possible to reduce the influence of friction or the like that occurs each time at a specific position.

 Note that T (N / 2 + 1) -T (N / 2) and the like can be used as the numerator of the above formula.

 Alternatively, T (N / 2) or {T1 + T2... T (N / 2)} × 2 / N may be used as the denominator.

Since the above calculation of the roughness degree RO is performed each time the required time T is measured, that is, every time the REF signal is output, the cylinder number n
Is defined as the roughness degree RO (n) of n cylinders. Therefore, the roughness degree RO (n) of each cylinder is sequentially obtained according to the ignition order.

Here, the roughness degree RO is 0 assuming that there is no angular velocity fluctuation. If there is a misfire, the roughness degree RO of a certain cylinder corresponding to the misfire indicates a negative value, and the roughness degree RO of another cylinder indicates a relatively positive value. In addition, roughness degree RO
The cylinder number n in (n) does not always match the misfiring cylinder.

Next, a moving average TRAVLU of the absolute value is calculated using the roughness degree RO (n).

That is, Is performed for each REF signal (step 5).

In addition, the above TRAVLU is set to the past maximum value during the data collection period.
Compare with TRAVMX, and if it exceeds, update the memory contents as a new maximum value TRAMMX (step 6). This is the final result of one data collection period
This indicates the maximum value of TRAVLU.

Then, in step 7, an index FU (n) indicating the positive / negative regularity of each cylinder n is calculated. This is FU (1)-FU (N)
Are indicated for each cylinder by N counters. If the roughness degree RO (n) of a certain cylinder n is positive, FU (n)
Is incremented, and if negative, FU (n) is decremented. Therefore, when misfire repeatedly occurs in a certain cylinder n, the index FU (n) of that cylinder n greatly decreases to the negative side.

Next, at step 8, the missing teeth of the ring gear 21 are detected. This will be described later.

In step 9, the moving average TR obtained in step 5 above
The AVLU is compared with the reference value TRAVCO. If the AVLU is equal to or larger than the reference value, the value of the first parameter MMF1 is incremented (step 10). The above reference value TRAVCO is the engine speed Ne.
Alternatively, it is given as a moving average of a reference value COMPL that is sequentially looked up from a map of the engine speed Ne and the basic fuel injection amount Tp.

Since TRAVLU is the average of the absolute values of the roughness degree RO as described above, the first parameter MMF1 is related to the frequency of angular velocity fluctuation of a certain level or more for the entire engine.

In step 11, it is determined whether or not a predetermined data collection period has elapsed. If the data collection period has elapsed, the process proceeds to step 12, and a failure determination is performed. The data collection period is defined by time or crank angle, and for example, about 2 to 3 seconds is sufficient. Therefore,
Until the predetermined data collection period elapses, the above-described processing of steps 3 to 10 is repeated, and the calculation of TRAVLU and the like are performed. Then, after performing the failure determination, TRAV
MX and the like are cleared to prepare for the next determination (step 13).

FIG. 8 shows details of the failure determination in step 12 described above. In step 41, a second parameter MMF2 and a third parameter MMF3 are calculated.

The second parameter MMF2 is obtained as MMF2 = TRAVMX / TRAVCO using TRAVMX, which is the maximum value of TRAVLU during the data collection period, and the above-described reference value TRAVCO. The value at the end of data collection is used for TRAVCO. The second parameter MMF2 indicates whether a large angular velocity fluctuation has occurred.

The third parameter MMF3 is the sum of all cylinders of the absolute value of the index FU (n) indicating the above-mentioned positive / negative regularity of each cylinder, that is, MMF3 = | FU (1) | + | FU (2) | ... + | FU (N) | Therefore, if the angular velocity decreases or increases in the same cylinder a plurality of times with the same tendency, the third parameter MMF3 increases. If the decrease or increase in the angular velocity appears at random, the value of the third parameter MMF3 does not increase so much.

In step 42, the value of the first parameter MMF1 counted in step 10 described above is compared with a reference value JMF1. If the value is less than the reference value JMF1, the process proceeds to step 43 and the second
The value of the parameter MMF2 is compared with the reference value JMF2. If the value is less than the reference value JMF2, the process proceeds to step 49, where it is determined that "normal", that is, there is almost no misfire. Meanwhile, step 43
If the value of the second parameter MMF2 is greater than or equal to the reference value JMF2, it means that although the angular velocity fluctuation is small as a whole engine, but large angular velocity fluctuation sometimes appears, the process proceeds to step 50, and the occurrence of low frequency misfire occurs. Judge as the state.

If the first parameter MMF1 is equal to or larger than the reference value JMF1 in step 42, it means that the angular velocity fluctuation occurs frequently in the entire engine, and thus the third parameter MMF3 is changed to the first reference value JMF3H and the second reference value JMF3H. JMF3L (however, JMF3H> JMF3L)
Are sequentially compared (steps 44 and 45). In the case of JMF3H or more, it means that the decrease or increase of the angular velocity repeatedly appears in the same cylinder with the same tendency as described above, and therefore, it is determined that the frequent misfire of one specific cylinder has occurred (step 46) And the misfiring cylinder is determined (step 47). The misfiring cylinder can be determined based on the index FU (n). In the present embodiment, the ignition cylinder immediately before the cylinder n having the smallest FU (n) corresponds to the misfiring cylinder.

If the third parameter MMF3 is less than JMF3H and less than JMF3L, it means that large angular velocity fluctuations appear frequently and randomly, and it is determined that a frequent misfire state of a plurality of cylinders has occurred (step 48).

Through the above processing, a detailed determination of the misfire state can be made. When it is determined that a misfire has occurred, the alarm lamp 24 is turned on.

Next, the flowchart of FIG.
The details of the gear loss detection process will be described. That is, if the teeth of the ring gear 21 are missing in the portion corresponding to the predetermined measurement period, the detected value of the required time T described above becomes abnormal only in that portion, but assuming that there is no change in the angular velocity, Assuming that the number of measured teeth in m is m, Is established.

Therefore, in this process, first, at step 31, RO (n)
/ TRAVLU (Α is an appropriate minute value) is determined, and if it is within the range, the counter CRGNG is sequentially incremented (step 33).

Then, the value of the CRGNG is compared with a predetermined reference value CRGNG1 (step 34), and if it is equal to or more than the reference value CRGNG1, it is determined that the ring gear is defective (step 35). If the value is outside the above range, CRGNG is set to 0 (step 32).

Therefore, the missing teeth of the ring gear 21 can be reliably detected,
The misjudgment of misfire due to this can be prevented. Here, as the processing after the detection of the gear loss, for example, the misfire determination may be stopped, and an alarm lamp may be turned on for warning, or the number of measured teeth at the corresponding location may be reduced. The correction may be made as m-1.

Incidentally, it is also possible to detect missing teeth of two, three, etc. in the same manner.

An embodiment of the present invention has been described above. Next, an embodiment will be described in which the processing accuracy and the like of the ring gear 21 are learned and corrected.

FIG. 9 shows the main flowchart.
After calculating the roughness degree RO (n) for each cylinder (n) (step 4), a roughness degree learning value LRNRO (n) is calculated (step 14). In other words, the roughness degree RO is given by
The moving average for each cylinder in (n) is calculated as a weighted average.

LRNRO (n) = W × RO (n) + (1−W) × LRNRO (n) (W is a very small weighting factor) The roughness degree learning value LRNRO (n) is, for example, the engine speed Ne or Ne. -Stored in the form assigned to the Tp map.

Then, in the calculation of the moving average TRAVLU in the next step 5, the moving average of the absolute value of {RO (n) −LRNRO (n)} obtained by subtracting the roughness learning value LRNRO (n) is obtained. Similarly, the calculation of the index FU (n) in step 7 is performed based on the sign of {RO (n) -LRNRO (n)}.

Therefore, when the required time T in each cylinder is slightly different due to the processing accuracy of the ring gear 21 or the like,
The shift is learned and corrected, and the misfire determination can be performed with higher accuracy.

EFFECT OF THE INVENTION As is apparent from the above description, according to the misfire detection apparatus for an internal combustion engine according to the present invention, an angular velocity variation is detected based on the number of signals output corresponding to a predetermined crank angle. In addition, since it is determined whether or not a misfire has occurred based on the tendency of angular velocity fluctuation during a certain length of time, a very high accuracy misfire detection can be performed. Also,
Based on the first to third parameters, a frequent misfire occurrence state in one cylinder, a frequent misfire occurrence state in a plurality of cylinders, and a low-frequency misfire occurrence state can be distinguished. Therefore, when a failure occurs, the cause of the failure can be easily determined, and accurate maintenance and repair can be performed in a short time.

In addition, by adding the learning correction, when a ring gear is used as a sensor, it is possible to eliminate the influence of the processing accuracy of the ring gear.

[Brief description of the drawings]

FIG. 1 is a claim correspondence diagram showing a configuration of the present invention, FIG. 2 is a claim correspondence diagram showing a configuration of the invention with learning correction added thereto, and FIG. 3 is a configuration explanation showing a mechanical configuration of an embodiment of the present invention. FIG. 4 is a time chart for explaining the operation of this embodiment. FIGS. 5, 6, 7, and 8 are flowcharts showing an example of the processing in this embodiment. 9 is a flowchart illustrating an example of performing learning correction. DESCRIPTION OF SYMBOLS 1 ... Reference position detection means, 2 ... Sensor, 3 ... Required time measurement means, 4 ... Roughness degree calculation means, 5 ... First parameter calculation means, 6 ... Second parameter calculation means, 7
... Third parameter calculation means, 8... Misfire state identification means, 9... Learning value calculation means, 10.

Continuation of front page (72) Inventor Shinhei Nakaba 1671-1, Kasukawa-cho, Isesaki-shi, Gunma Japan Electronic Equipment Co., Ltd. (56) References JP-A-3-206337 (JP, A) (58) Fields investigated (Int) .Cl. ⁶ , DB name) F02D 45/00

Claims (2)

(57) [Claims] 1. A reference position detecting means for outputting a reference position signal at a reference crank angle position of each cylinder, a sensor for outputting a predetermined signal at each predetermined crank angle, and a predetermined phase based on the reference position signal. Means for measuring a required time required to detect a predetermined number of signals output from the sensor, and means for calculating a degree of roughness of each cylinder by dividing a deviation of the sequentially measured required time by an average value of the required time. Means for calculating the first parameter by counting the number of times the average value of the absolute value of the roughness degree of all cylinders exceeds a predetermined reference value, and calculating the maximum value of the absolute value of the roughness degree within a predetermined period. Means for obtaining a second parameter based on the first and third parameters, means for adding and subtracting the number of positive and negative of the roughness degree of each cylinder for a predetermined period, and obtaining a third parameter as the sum of the absolute values thereof. Hazuki misfire detecting device for an internal combustion engine, characterized by comprising means for identifying and frequent misfiring state of misfire state and a low frequency at frequent misfiring state and a plurality of cylinders of one cylinder. Means for calculating a roughness degree learning value as a moving average for each cylinder for each calculation of the roughness degree; and means for correcting the sequentially calculated roughness degree using the roughness degree learning value. 2. The misfire detection device for an internal combustion engine according to claim 1, wherein: