WO1997020195A2

WO1997020195A2 - Process for detecting misfires in an internal combustion engine

Info

Publication number: WO1997020195A2
Application number: PCT/DE1996/002097
Authority: WO
Inventors: Michael Henn; Anton Angermaier
Original assignee: Siemens Aktiengesellschaft
Priority date: 1995-11-30
Filing date: 1996-11-04
Publication date: 1997-06-05
Also published as: DE19544720C1; WO1997020195A3

Abstract

A misfire detection process based on the crankshaft speed of rotation is disclosed. An unsteady running value is derived from measured cylinder segment times and subsequent correction of mechanical teeth defects. By a cylinder-selective disturbance variable compensation dependent on load and speed of rotation, the disturbance caused by torsional oscillations of the crankshaft is taken into account.

Description

description

Method for detecting misfires in an internal combustion engine

The invention relates to a method for detecting misfires in an internal combustion engine by evaluating the crankshaft speed according to the preamble of patent claim 1.

The occurrence of misfires in an internal combustion engine can, on the one hand, lead to an increase in the emission rate of pollutants and, on the other hand, due to after-reactions of the unburned air / fuel mixture, to the destruction of the catalytic converter arranged in the exhaust tract of the internal combustion engine or at least to an impairment of its conversion ability.

The detection of such misfires is therefore required in order to monitor compliance with the legal limit values for emissions during operation. Detection of the misfires from the crankshaft speed measured by means of incremental encoders offers a cost-effective implementation.

A large number of methods are therefore already known which measure segment times for the detection of misfires in combustion which the crankshaft requires during the work cycles of the individual cylinders to pass through predetermined angular ranges. Then, uneven running values are calculated from the segment times and these values are compared with threshold values, errors in the segment time measurement being recognized and corrected in times of the overrun fuel cutoff of the internal combustion engine (z3. EP 0 583 496 AI). Misfires lead to a temporary slowdown of the angular velocity of the crankshaft, this effect being very slight, so that the angular velocity must be determined with high precision.

EP 0 576 705 A1 therefore proposes, in a method for detecting misfires, to take into account, in addition to a static component, the general trend in speed and additionally non-uniform changes in speed when measuring the speed, so that even in highly unsteady operation, fault detections are largely ruled out are.

Tolerances and manufacturing tolerances in the production (eg. Mechanical tooth error) or during assembly of the incremental encoder on the crankshaft (eg. Exentrische storage) also the Winkelge ^¬ lead to inaccuracies in the measurement speed, and thus to possible false detections in the misfire detection.

EP 0 583 495 A1 discloses a method for detecting and correcting errors in the time measurement on rotating shafts, in particular on crankshafts. In this case, segment times are measured which the shaft needs in order to rotate a defined angular range (segment) and then these times are compared with a time which applies to a reference segment. In the trailing operation of the internal combustion engine, either a correction value is determined which enables the measured segment time to be corrected individually for each cylinder or for each segment.

In addition, reactions from the road and the mechanical behavior of the crankshaft itself influence the speed behavior of the crankshaft and thus make it more difficult to detect combustion misfires. Especially with multi-cylinder Three third in-line engines with a long crankshaft no longer neglectable torsional vibrations, which can falsify the result of misfire detection. This can go so far that in higher speed ranges, the detection reliability can no longer be guaranteed for all cylinders with conventional methods.

The invention is based on the object of specifying a method which, compared to the known prior art, enables even more precise and reliable detection of combustion misfires.

According to the invention, this object is achieved by the features of patent claim 1. Advantageous embodiments of the invention are characterized in the subclaims.

An uneven running value is determined from the measured cylinder segment times after a correction of the mechanical tooth errors which are periodic over one revolution. By means of a cylinder-selective gate size connection depending on

The load and speed range of the operating point of the internal combustion engine suppresses the disturbance caused by the torsional vibrations of the crankshaft. The values for the store size are determined on a vehicle test stand and stored in cylinder-specific maps.

An embodiment of the invention is explained in more detail below with reference to the schematic drawings. Show it:

FIG. 1: a schematic representation of the measuring principle for determining the angular velocity of a crankshaft, FIG. 2: diagrams for the differences in the segment times without correction, FIG. 3: segment time differences with correction by subtraction, FIG. 4: segment time differences with correction by subtraction and multiplication, and Comparison of the uneven running values with and without correction of the torsional moment for measured values.

In FIG. 1, the reference numeral 1 designates a transmitter gearwheel having ferromagnetic teeth with winding increments of width Δφ and is mounted on a crankshaft 2. From a magnetic pickup 3, e.g. a Hall sensor or an inductive sensor, a voltage signal is generated during the rotational movement of the crankshaft 2, which fluctuates with the distance of the gear face. The gearwheel thus forms the modulator for transforming the amplitude-analog input high angular velocity m into a frequency-analog signal. The zero crossings of this signal also contain the information about the current angle. The sequence of the tooth gaps and the ferromagnetic tooth of the gear wheel 1 changes the magnetic field that comes from a permanent magnet in the sensor 3.

From the signal supplied by the sensor 3, a discriminator 4, which can consist, for example, of a Schmitt trigger and a edge detector, generates a square-wave signal, which is referred to as segment time by the distance between two edges T _n , hereinafter is marked. This signal is quantized using a counter 5 and a reference

Aφ frequency 6. The counter number obtained in this way is εm

T (n) for the angular velocity ω. By omitting one or more teeth on the encoder gear 1, a region 7 is obtained for an angle reference, with the aid of which the absolute angle can be determined. 60 teeth minus a gap of 2 teeth have established themselves as the standard for pulse generators on the crankshaft of internal combustion engines.

If the angular velocity of the crankshaft is determined by means of an incremental encoder, the time T _n for sweeping over a cylinder segment is measured, for example in the case of a 6-cylinder, four-stroke internal combustion engine. above 120 °, the value sequence of T _{n can be used} to detect misfires.

In the difference between two successive segment times ΔT _n = T _n -T _n _ ₎ or the segment time difference standardized with Tl

AT _n / T *, a misfire then manifests itself in a characteristic dip in the signal curve. Unfortunately, even with stationary operation of the internal combustion engine without misfires, ie in the event of a fault, the consequences .DELTA.T _n or AT _n IT _{n have} cylinder-specific deviations that keep repeating at the same operating point for a specific vehicle. In high speed ranges in particular, these faults make reliable misfire detection difficult or even impossible. FIG. 2 illustrates this by way of example for a 4-cylinder

Internal combustion engines. In the left half of FIG. 2, the differences .DELTA.T _{n of} the segment times are plotted one after the other in accordance with the firing sequence 2-1-3-4, while in the illustration according to the right half of FIG. 2 they are split up according to cylinders .

The causes of these faults lie on the one hand in the manufacturing tolerances of the transmitter gear and on the other hand in the Torsional vibrations of the crankshaft. While the mechanical inaccuracies of the transmitter gear have to be determined for each individual vehicle in an adaptation process, the torsional disturbances are characteristic of a particular vehicle type. To eliminate the influence of the interference, the two sources of interference can be separated by adapting the gearwheel errors in operating areas in which there are little or no torsional disturbances. The deviations in the segment times of the individual cylinders, due to the torsion of the crankshaft for a vehicle type, can be determined from the measurements in error-free, stationary operation on a test bench from the values ΔT _n or AT _n IT _n corrected for the gearwheel errors or from values derived therefrom become. If the data are stored in a map for each cylinder over the load and the speed of the internal combustion engine, interference size compensation can be carried out by subtracting the values interpolated from the map

If the signal profiles according to FIG. 2 are corrected in this way, the result is a representation as indicated in FIG. 3.

The subtraction of the disturbance variable pulls the values for all cylinders to zero level on average only when the internal combustion engine is operating without interruptions. If a misfire occurs, the signal drops differ for the segment time differences of different cylinders. The multiplication by a factor zylindermdividuellen leveled this un ^¬ terschiedlichen heights in dropouts case. The factors can be determined on a chassis dynamometer at dropouts operation of the internal combustion engine ^¬ and stored in a map. FIG. 4 shows the signal curve .DELTA.T _n of the cylinders assimilated by the additional multiplication In the method according to EP 0 576 705 A1, the measured segment times T _{n are used to} calculate an uneven running value for each cylinder

LU _{n is} calculated on the basis of the difference ΔT _{n between} two cylinder segments. This is compared with a threshold LUG _n , which are calculated using characteristic maps from the measured load, speed and temperature. If the threshold is exceeded, a dropout is then decided. The method described in EP 0583 496 AI also takes mechanical gear tolerances into account by correcting the measured segment times T _n to TK _n : τκ _n = (ι-κ _n ) τ _n .

The correction factors K _n are adapted in the overrun fuel cut-off mode of the engine.

If such a method of uneven running is used without taking into account the disturbance caused by torsional vibrations, there are large differences in the uneven running values for the individual cylinders with increasing speeds. The low signal swing that a misfire generates at high speeds and low loads can lead to misdetection or misrecognition of misfires.

The torsional vibrations of the crankshaft are suppressed in the following with a disturbance variable. In the case of a flow diagram of an algorithm for misfire detection, there are various options for taking such a feed forward into account:

• The correction can already be carried out together with the correction for mechanical gear tolerances:

TK _n = (lK _n -KTOR _n ) T _n The correction factors KTOR _n , which contain the influence of the torsional moment, are calculated from cylinder-specific characteristic diagrams as a function of load and speed.

• The next possibility is to subtract a cylinder-specific value LUTOR _n when calculating the rough running:

LU _n = LU ^* -LUTOR _n

LUTOR _n is in turn determined from load and speed-dependent maps for each cylinder. LU _rl is the calculated uneven running without taking the torsional vibration into account

• Finally, you can also take into account the disturbance variable feed-in by means of cylinder-specific characteristic maps when determining the threshold value LUG _n .

A method for misfire detection in which the current torque is taken into account by means of cylinder-specific correction values is now explained in more detail with reference to FIG.

In a first method step SI, the segment times T _{n are} measured, ie the time periods that the crankshaft needs in order to rotate through a certain crank angle during the work cycle of a cylinder. This can be done, for example, with a device according to FIG. 1. The measured segment times T _n are then corrected in method step S2. Such a correction is necessary, since the inaccuracies which occur due to tolerances and specimen variations during manufacture or when the incremental angle sensor, for example a sensor wheel, on the crankshaft are installed, lead to a faulty detection. averaging of the angular velocity and thus possible misdetection of misfires. The tooth error correction in step S2 takes place according to the relationship τκ _n = (ι -κ _n ) τ _n with T _n : measured, uncorrected segment time K _n : correction factor

Tk _n: corrected segment time

The correction factors K _n are adapted in the operating state of the overrun cutoff (towing mode) of the internal combustion engine, as described, for example, in EP 0 583 495 AI. In this known method for the detection and correction of errors in the time measurement of rotating parts, the segment time of a reference segment of a reference cylinder is measured, stored and then the segment times of the segments belonging to the individual cylinders are measured successively for all cylinders. For the same reference cylinder, two crankshaft revolutions later, the segment time of this reference segment is measured, and for the individual segments assigned to the cylinders of the internal combustion engine, successive correction values are calculated and these are then averaged.

From the segment times TK _n corrected in this way, before taking the mechanical gear wheel tolerances into account, so-called uncompensated runs

_* resting values LU _n calculated according to any known method. Depending on the type of method used, various dynamic influences that occur during operation of the internal combustion engine (acceleration, deceleration) are compensated for. A possible method for calculating the current values is, for example, in the European Tent application EP 0 576 705 AI described. According to this known method, the uneven running value is determined for each cylinder by measuring the successive time periods (segment times) which the crankshaft needs during the work cycles of the cylinders which follow one another with regard to the firing sequence in order to run through predetermined angular ranges. These rough running values are composed of a static component, a dynamic component that takes into account the general trend in speed, and a component that takes into account changes in acceleration and deceleration. These proportions are calculated on the basis of the difference in the time periods of directly successive cylinders or on the basis of the difference in the time periods of cylinders which are further apart.

Simultaneously with the measurement of the segment times T _n , the speed, the load and the temperature of the internal combustion engine are continuously measured in process step S4 via appropriate sensors. In method step S5, the additive disturbance variables LUTOR _n , which take the torsional vibration into account, are determined from characteristic maps spanned over the speed and load and stored in a memory of an electronic control device of the internal combustion engine. There is a separate map for each cylinder. The selection is made by means of a cylinder identification signal ZYL_ED.

This signal can be obtained, for example, by determining the absolute angle of the crankshaft with the help of the tooth gap of the transmitter wheel as an angle reference and by em signaling a camshaft transmitter wheel.

The values stored in the characteristic diagrams are determined on a vehicle test stand for the corresponding drive type. In method step S7 the disturbance variable LUTOR _n is subtracted from the uncompensated uneven running value LU _n from method step S3 in order to obtain the compensated uneven running value LU _n . This compensated uneven running value LU _n is compared in method step S8 with a threshold value LUG _n , which are calculated using characteristic maps from the measured load, speed and temperature of the internal combustion engine (method step S7).

If the uneven running value with disturbance variable input LU _{n is} smaller than the threshold value LUG _n , a misfire is registered in method step S9. If the value LU _n is greater than or equal to the threshold value, no misfire is registered (method step S10). Both results of the query mm method step S8 are fed to a statistical evaluation, since the risk of incorrect detection due to non-reproducible influences would be too great for individual combustion misfires detected. Control measures, such as switching off the injection to individual cylinders, are therefore only taken when the statistical frequency of such misfires exceeds a certain, predetermined limit value.

It turns out that the rough running values for a misfire for different cylinders can be slightly different, with otherwise the same operating conditions of the internal combustion engine. This can be explained by the fact that the disturbance due to the torsional vibration in the case of a misfiring varies slightly compared to the faultless operation without misfires. If necessary, a factor GFAK _n can also be taken into account for approximate correction when calculating the compensated uneven running value LU _n in step S7: LU _n = GFAX _n (LU ^* - LUTOR _n ) (57a)

GFAK _n is an individual weight factor that is determined by a map of the engine speed.

The results obtained with this method for a 6-cylinder in-line engine at a speed of 6000 1 / min and a load of 250 mg / stroke are shown in the diagram in FIG. 6. In the upper half of this figure the calculated uneven running values LU _n plotted at certain discrete angles n without taking the torsion correction into account.

The lower half of FIG. 6 shows the rough running values LU _n with torsion correction according to the method according to the invention for the same internal combustion engine at the same operating point. If the two representations are compared with one another, it can be seen that the risk of incorrect detection of misfires in combustion can be significantly reduced by taking into account the disturbance variable torsional torque. If, for example, the threshold value LUG _n , when it is detected that a misfire is detected, is set to the value -50, then uneven running values above the threshold value are erroneously registered as combustion misfires without correction, even though the over - Stepping does not result from a misfire, but occurs due to torsional vibrations of the crankshaft.

Claims

claims

1. Method for detecting misfires in a multi-cylinder internal combustion engine by evaluating the crankshaft speed with the following steps:

- Measuring the segment times (T _n ) that the crankshaft needs during the work cycles of the individual cylinders to run through predetermined angular ranges, - Correcting these segment times (T _n ) with a correction

factor (K _n ), which includes the mechanical tolerances of the third-party sensor,

- Calculation of rough running value (LU _n ) from the corrected segment times (TK _n ),

- Comparison of the uneven running values (LU _n ) with a threshold value (LUG _n ) and registration of a misfire when the threshold value is exceeded, characterized by an interference size dependent on the operating point of the internal combustion engine, which caused the torsional oscillations of the crankshaft Speed influence is taken into account.

2. The method according to claim 1, characterized in that the StorgroßenaufSchaltung by eme additional correction of

Segment times (T _n ) due to individual correction factors

ren (KTOR _n ).

3. The method according to claim 2, characterized in that the correction factors (KT0R _n ) depending on the load and the rotational number of internal combustion engines are stored in individual cylinder maps.

4. The method according to claim 2 or 3, characterized in that the interference size connection takes place according to the following relationship:

TK _n = (lK _n -KTOR _n ) T _n

5. The method according to claim 1, characterized in that the StorgroßenaufSchaltung by eme correction of the uneven running

_* π values (LU _n ) through cylinder-specific correction factors

(LUT0R _n ), which are stored as a function of the load and the speed of the internal combustion engine in characteristic diagrams specific to the cylinder.

6. The method according to claim 5, characterized in that the fault size connection is effected according to the following relationship:

LU _n = LU _n -LUTOR _n

7. The method according to claim 6, characterized in that in the case of interference size addition, an additional weight factor (GFAK _n ) according to the relationship

LU _n = GFAK _n (LU _n -LUTOR _n l is taken into account, which is stored in a map as a function of speed.

8. The method according to claim 1, characterized in that the interference size activation by cylinder-specific threshold values (LUG _n ) for the Laufunru¬

values (LU _n ).