Method for determining the angular position of an engine

Abstract

A method for determining the angular position of an engine by a crankshaft sensor, having the following steps: production by the crankshaft sensor of a signal exhibiting a revolution event, determination of the revolution out of two revolutions, since a crankshaft makes two revolutions per engine cycle for a four-stroke engine, for each no tooth event potentially produced, a change in the direction of rotation of the engine is suspected, and an analysis step is performed which comprises: if, during an inverse window, a further no tooth event is produced, the change in direction of rotation is confirmed.

Claims

1. A method for determining an angular position of an engine by a crankshaft sensor comprising a crankshaft detector facing a crankshaft toothed wheel comprising a large number of regular teeth and a revolution marker, the crankshaft detector being able to produce a signal having a tooth event corresponding to an edge for each of the teeth, a revolution event for the revolution marker, and a no tooth event when two successive tooth events are abnormally separated, the method comprising: producing, by the crankshaft sensor, a signal having the revolution event, determination of the revolution event out of two revolutions for a four-stroke engine, a crankshaft performing exactly two revolutions per cycle of the engine, in order to complete the determination of the angular position of the engine, wherein for each of the no tooth events possibly produced, a change in direction of rotation of the engine is suspected, and an analysis step is carried out comprising: a) if in an inverse window, at a distance from a current no tooth event equal to a distance between the preceding revolution event and the current no tooth event, toleranced by +/a tolerance of the teeth, a new no tooth event is produced, the change in direction of rotation is confirmed, and b) if in the inverse window no no tooth event is produced, the change in direction of rotation is invalidated.

2. The method as claimed in claim 1, in which the analysis step additionally comprises: c) if in a direct window, at a distance from a preceding revolution event equal to a crankshaft wheel revolution, 2olerance by +/the tolerance of teeth, a new revolution event is produced, the change in direction of rotation is invalidated, and d) if in the direct window no new revolution event is produced, the change in direction of rotation is confirmed.

3. The method as claimed in claim 1, wherein the no tooth event can be produced only outside a direct window at a distance s from a preceding revolution event by a tooth event number equal to the large number of teeth corresponding to a rotation of the crankshaft toothed wheel and tolerance by +/the tolerance of teeth, wherein the tolerance is equal to 2 teeth.

4. The method as claimed in claim 1, wherein the revolution event can be produced only in a direct window at a distance from a preceding revolution event by a tooth event number equal to the large number of teeth corresponding to a rotation of the crankshaft toothed wheel and tolerance by +/the tolerance of teeth, wherein the tolerance is equal to 2 teeth.

5. The method as claimed in claim 1, wherein the crankshaft toothed wheel is regularly angularly divided into 60 and the large number of teeth is equal to 58, and 2 consecutive missing teeth form the revolution marker.

6. The method as claimed in claim 1, wherein the tolerance is equal to 2 teeth.

7. The method as claimed in claim 2, wherein the no tooth event can be produced only outside a direct window distant from the preceding revolution event by a tooth event number equal to the large number of teeth and tolerance by +/the tolerance of teeth, wherein the tolerance is equal to 2 teeth.

8. The method as claimed in claim 1, wherein each of conditions a) and b) are considered in order of occurrence.

9. The method as claimed in claim 2, wherein each of conditions c) and d) are considered in order of occurrence.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Other features, details and advantages of an aspect of the invention will become more clearly apparent from the detailed description provided below by way of indication with reference to the drawings, in which:

(2) FIG. 1 shows, on a timing diagram, a crankshaft signal over one complete engine cycle,

(3) FIGS. 2-6 show, in a time diagram, a crankshaft signal according to different use cases, as follows:

(4) FIGS. 2 and 3 show two use cases having a no tooth event situated in a first half of a revolution,

(5) FIGS. 4 and 5 show two use cases having a no tooth event situated in a second half of a revolution,

(6) FIG. 6 shows another use case having a second no tooth event.

(7) The crankshaft is the output shaft of an engine. It rotates driven directly by the connecting rod or rods and performs two revolutions per engine cycle. A camshaft, controlling the valves, is a shaft driven indirectly, via a distribution transmission, by the crankshaft, and performs one revolution per engine cycle. An engine cycle is then conventionally labeled as a function of the angle of orientation of the crankshaft over 720.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

(8) A crankshaft or CRK sensor makes it possible to know the angular position of the crankshaft. To this end, a crankshaft sensor comprises a crankshaft toothed wheel and a crankshaft detector arranged facing said crankshaft toothed wheel and able to detect a presence/absence of material and thus to detect a tooth or a slot. The crankshaft toothed wheel is joined to the crankshaft so as to rotate therewith, whereas the crankshaft detector is fixed. The crankshaft toothed wheel comprises a large number N of regular teeth and a single revolution marker that makes it possible to determine a particular angular position upon each revolution in an absolute fashion. The crankshaft toothed wheel is angularly divided equally into a large number of regular teeth, thus making it possible to accurately ascertain the angular position of the crankshaft, by counting the teeth, in relation to the revolution marker. Said revolution marker is generally associated with a particular position of the engine, such as conventionally the top dead center of a cylinder, for example the first cylinder.

(9) The crankshaft sensor arranged facing the crankshaft toothed wheel is able to detect a presence of material facing a tooth and an absence of material facing a recess or slot. The crankshaft detector or a processing unit, which is associated and merged with the crankshaft detector for the needs of the present case, is able to produce a tooth event d for each of the N teeth of the crankshaft toothed wheel. Such a tooth event d typically corresponding to an edge for each tooth. Given the large number N of teeth present on the crankshaft toothed wheel, a single edge per tooth, out of the rising edge or the falling edge, may be adopted. In a conventional manner, the falling edge is used to form the tooth event d. This hypothesis is adopted for the remainder of the description.

(10) The crankshaft detector is additionally able to produce a revolution event T when it detects the revolution marker.

(11) The profile of the teeth of the crankshaft toothed wheel is symmetrical. It therefore does not make it possible to ascertain the direction of rotation of the crankshaft toothed wheel and of the crankshaft. The direction of rotation of the engine, and therefore of the crankshaft, is assumed to be normal, initially on starting, when the synchronization method is implemented. However, this direction of rotation may be inverted in certain circumstances, causing the engine to rotate in the inverse direction.

(12) To simplify the description, it is assumed that the tooth events d are produced on falling edges. An identical reasoning could be applied for rising edges.

(13) At the moment of the inversion of the direction of rotation, the crankshaft detector sees a last falling edge, since the tooth events d are falling edges, then a last recess where the stopping of the rotation occurs according to a first hypothesis. Alternatively, according to a second hypothesis, the rotation continues and the crankshaft detector sees another last rising edge, hence ignored since rising, preceding a last tooth where the stopping of the rotation occurs.

(14) When the crankshaft toothed wheel sets off in the other direction, according to the first hypothesis the crankshaft detector sees, in the other direction, the start of the last recess. It then sees a rising edge, hence ignored since rising, which is other than the last falling edge seen in the other direction. It then sees a tooth and a falling edge, which forms a new tooth event d.

(15) When the crankshaft toothed wheel sets off in the other direction, according to the second hypothesis the crankshaft detector sees, in the other direction, the start of the last tooth. It then sees a falling edge, which forms a new tooth event d. This falling edge is other than the last rising, edge seen in the other direction.

(16) The result of this is that the last rising edge seen before the change of direction and the first following falling edge seen after the change of direction produce tooth events d which are most often closer to or more distant from one another than two tooth events d produced by two successive teeth seen in one and the same direction of rotation. Such a variation in the distance/periodicity between two successive tooth events d during a change of direction, by comparison with a previous distance/periodicity in one and the same direction of rotation, can be identified by the crankshaft detector which consequently produces a no tooth event DD or a revolution event.

(17) Certain processing algorithms make it possible to avoid a confusion between a revolution event T and a no tooth event DD, mainly on the basis of the periodicity of the revolution events T.

(18) According to one common but non-mandatory embodiment, the crankshaft toothed wheel is angularly divided equally into 60 regular teeth. Two consecutive teeth are removed so as to form the revolution marker. This leads to a signal CRK as seen by the crankshaft detector, as illustrated in FIG. 1. The signal CRK periodically has a revolution event T at the 2 missing teeth and, more precisely, at the 1st tooth following the two missing teeth, followed by 58 tooth events d, as long as the crankshaft is rotating in one and the same direction.

(19) According to certain implementations, a revolution event coincides with a first tooth event and thus occults the latter. Thus, the following tooth events theoretically numbering 58 are 57 in number in this particular practical case.

(20) Following detection of a revolution event T1, a new revolution event T2 is expected, in a direct window F2, one revolution of the crankshaft toothed wheel later. It is advantageously verified that this new revolution event T is situated in a window of N=58+/ n=2 tooth events d (including, where appropriate, the tooth event coinciding with the revolution event) after the preceding revolution event T1.

(21) In order to avoid confusing a revolution event T with a no tooth event DD, a similar window of N=58+/ n=2 tooth events d is employed after each revolution event T, in which it is not possible to produce a no tooth event DD even if a new revolution event T can be produced only in this window of N=58+/ n=2 tooth events d after each preceding revolution event T.

(22) As soon as a first revolution event T is detected, the angular position of the crankshaft toothed wheel, and therefore of the crankshaft, is known with an inverse precision of the total number of teeth N+2, including the two missing teeth, of the crankshaft toothed wheel, that is to say all the more precise as the number N of effective teeth or the total number N+2 of teeth is large. The crankshaft is synchronized. It is therefore advantageous for the crankshaft toothed wheel to comprise a large number N of teeth.

(23) However, for a four-stroke engine, a crankshaft performs exactly two revolutions per engine cycle. Thus, the knowledge of the angular position of the revolution marker and the synchronization of the crankshaft are insufficient to indicate the angular position of the engine, since it is known with an uncertainty of one revolution out of two.

(24) The determination of the revolution out of two, in order to complete the determination of the angular position of the engine, can be carried out by any means. This point is not the subject of an aspect of the invention. According to one embodiment, it is possible to use a camshaft sensor, for example according to the method as described in patent application FR 1560189 of 26 Oct. 2015 (published as FR 3042860) by the same applicant, incorporated by reference herein.

(25) It is always assumed that the engine initially rotates in the normal direction.

(26) Outside the tooth events, the first event produced by the crankshaft sensor is always a revolution event T, denoted T1.

(27) Specifically, any anomaly, whether it concerns a revolution marker, a sudden acceleration, or a change in direction of rotation, will be detected in the same way. Thus, according to one possible embodiment, an anomaly is detected, for example, by means of a comparison of the successive tooth distances. This can, for example, be implemented by a formula: Td(i)/(Td(i1)>K, where Td (i) is the duration of the ith tooth between the front of a preceding tooth event i1 and the front of a following tooth event i, and K is a detection threshold, typically equal to 1.5. In the nominal case of normally spaced teeth, the ratio is substantially close to 1. If inequality is verified, with a ratio above K, an anomaly is detected. This test is a possible means for determining that two successive tooth events d are abnormally separated.

(28) The very first anomaly thus detected is considered to be a revolution marker. This hypothesis may, where appropriate, be verified by means of a stricter formula than the preceding one. If a revolution marker is confirmed, a first revolution event T, denoted T1, is produced.

(29) Once this first revolution event T1 is produced, there is periodically determined a direct window F1, F2 in which a new revolution event is expected. This direct window F1, F2 is determined to be distant from the first revolution event T1 by a crankshaft wheel revolution, that is to say by the tooth number N of the crankshaft wheel assigned a tolerance of +/ n teeth. Thus, any new anomaly produces a new revolution event T2 if it is situated in such a direct window F1, F2 or a no tooth event if it is situated outside such a direct window F1, F2.

(30) Such a method of synchronizing/determining the angular position of an engine can, in a prejudicial manner, be deceived, for example, if the engine changes direction of rotation and starts to rotate oppositely. Now, if an engine is said to be synchronized, whereas it rotates oppositely, a damaging operation, such as fuel injection, can be commanded and can lead to damaging effects for the engine.

(31) An inversion or change in direction of rotation of the engine is necessarily accompanied by a no tooth event DD which is assumed to be always detectable by the crankshaft sensor. However, a no tooth event DD can also be produced by other causes. Thus, it is appropriate to know the difference in order to confirm or invalidate a change in direction of rotation of the engine.

(32) In order to avoid such a problem, an aspect of the invention proposes that a no tooth event DD1, DD2 be considered as a suspicion of change in direction of rotation. Only one suspicion is taken into account since, given the mode of production of a no tooth event DD1, DD2, such an event can also be produced in the event of sudden deceleration of the engine, in the event of an engine sputter or else in the event of a very rapid back-and-forth change in direction of rotation, or double change in direction. In any case, the engine ultimately rotates in the normal direction and does not risk posing a problem for synchronization. By contrast, a revolution marker, seen outside a direct window F1, F2, typically owing to an inversion of direction of rotation, produces a no tooth event and not a revolution event.

(33) Such a suspicion of change in direction of rotation, triggered by a no tooth event DD1, also called current no tooth, must then be confirmed or invalidated, advantageously as quickly as possible.

(34) This is carried out by an analysis of the conditions or events occurring subsequently to the current no tooth event DD1, having raised the suspicion of change in direction of rotation.

(35) Two cases may present themselves: the current no tooth event DD1 is produced in the first half of a crankshaft wheel revolution, that is to say in the first half of the interval separating the preceding revolution event T1 from a following revolution event T2, or by contrast the current no tooth event DD1 is produced in the second half of the crankshaft wheel revolution, that is to say in the second half of the interval separating the preceding revolution event T1 from a following revolution event T2.

(36) The first case is illustrated in FIGS. 2 and 3. Since the current no tooth event DD1 is produced in the first half of a crankshaft wheel revolution, the closest event (outside a tooth event d) which can be produced subsequently is a new no tooth event DD2.

(37) If as illustrated in FIG. 2, this new no tooth event DD2 is situated in an inverse window I1, in that the distance between the preceding revolution event T1 and the current no tooth event DD1 is substantially equal to the distance between the current no tooth event DD1 and the new no tooth event DD2, these distances being represented by black arrows, the new no tooth event DD2 can be interpreted as an aliasing of the revolution marker previously having produced the preceding revolution event T1, and now seen (again) in the other direction. It then appears that the engine has probably changed direction of rotation and that the current no tooth DD1 indeed corresponded to a change in direction of rotation. Thus, if this condition of the presence of a no tooth event DD2 which is equidistant, or in an inverse window I1, is verified, the change in direction of rotation is confirmed.

(38) It should be noted that this confirmation of change in direction of rotation can be subsequently confirmed in that, since the engine is supposed to have changed direction of rotation, no revolution event should be produced in the next direct window F2.

(39) If, by contrast, as illustrated in FIG. 3, no new no tooth event is produced in the inverse window I1, situated at a distance from the current no tooth event DD1 substantially equal to the distance between the preceding revolution event T1 and the current no tooth event DD1, this condition can be interpreted as an absence of aliasing of the revolution marker. This condition includes the case of a new no tooth event DD2 produced but not situated in the inverse window I1, as illustrated in FIG. 6. It then appears that the engine has probably not changed direction of rotation and that the current no tooth DD1 corresponded to another cause, such as a sudden acceleration, and not a change in direction of rotation. Thus, if this condition of absence of a no tooth event DD in the inverse window I1 is verified, the change in direction of rotation is invalidated.

(40) It should be noted that this invalidation of change in direction of rotation can be subsequently confirmed in that, since the engine is not supposed to have changed direction of rotation, a new revolution event T2 should be produced in the next window F2.

(41) The second case is illustrated in FIGS. 4 and 5. Since the current no tooth event DD1 is produced in the second half of a crankshaft wheel revolution, the closest event (outside a tooth event d) which can be produced subsequently is a new revolution event T2. Specifically, any aliasing of the preceding revolution event T1 could be produced only after the direct window F2.

(42) If, as illustrated in FIG. 4, no new revolution event is produced in a direct window F2, distant from the preceding revolution event T1 by a crankshaft wheel revolution, that is to say substantially N tooth events d, this can be interpreted as the consequence of a change in direction of rotation of the engine. Thus, the current no tooth DD1 probably indeed corresponded to a change in direction of rotation. Thus, if this condition of absence of revolution event in a direct window F2 is verified, the change in direction of rotation is confirmed.

(43) It should be noted that this confirmation of change in direction of rotation can be subsequently confirmed in that, since the engine is supposed to have changed direction of rotation, an aliasing of the preceding revolution event T1 should produce a new equidistant no tooth event DD2, that is to say at a distance from the current no tooth event DD1 substantially equal to the distance between the preceding revolution event T1 and the current no tooth event DD1.

(44) If, by contrast, as illustrated in FIG. 5, a new revolution event T2 is produced in the direct window F2, situated at a distance from the preceding revolution event T1 substantially equal to a crankshaft wheel revolution, this condition can be interpreted as a confirmation that the engine still rotates in the normal direction. This condition includes the case of a new no tooth event DD2 produced but not situated in the window F2, as illustrated in FIG. 6. It then appears that the engine has, probably not changed direction of rotation and that the current no tooth DD1 corresponded to another cause, such as a sudden acceleration and not a change in direction of rotation. Thus, if this condition of the presence of a revolution event in the direct window F2 is verified, the change in direction of rotation is invalidated.

(45) It should be noted that this invalidation of change in direction of rotation can be subsequently confirmed in that, since the engine is not supposed to have changed direction of rotation, a new no tooth event DD2 should not be produced in a next inverse window I1, situated at a distance from the current no tooth event DD1 substantially equal to the distance between the preceding revolution event T1 and the current no tooth event DD1, confirming an absence of aliasing of the revolution marker.

(46) It should be noted that the case where the no tooth event is situated exactly in the center of the revolution, that is to say at equal distance from the preceding revolution event T1 and from the new revolution event T2, cannot be resolved by the method. Specifically, in this case, the direct window F2 coincides with the inverse window IL Thus, in this particular case, it is not possible to determine if an anomaly is a new revolution T2 (or no tooth) event or a revolution event caused by an aliasing of the revolution marker of the preceding revolution event T1, the two phenomena being superimposed.

(47) For this particular case, it is appropriate to employ another means for detecting a change in direction of rotation, such as that described in the aforementioned patent application. This other detection means can be used alternatively or additionally to the present invention.

(48) The previously described analysis has not to determine if a current no tooth event DD1 is in the first half or in the second half of the revolution. It is sufficient to apply the analysis by testing the following four conditions: presence or absence of a revolution event in the direct window F2, presence or absence of a no tooth event in the inverse window I1, and by reacting as a function of the condition which occurs first.

(49) The previously described analysis is advantageously applied to any no tooth event which can in turn be considered as a suspicion of change in direction of rotation. Thus, each successive no tooth event is advantageously considered as a current no tooth to which the preceding analysis is applied.

(50) Thus, as illustrated in FIG. 6, a first no tooth DD1 is produced. This no tooth DD1 is considered as the current no tooth and the analysis step is applied thereto by testing, where appropriate, the presence of another no tooth, such as, for example, the no tooth DD2. Next, a second no tooth DD2 is produced. This no tooth DD2 is in turn considered as the current no tooth, potential indicator of a change in direction of rotation, and the analysis step is applied thereto for the purpose of confirmation or invalidation. The procedure is thus for each successively produced no tooth event.

(51) As described above, the first detected anomaly is considered to be a revolution event. Next, a periodic direct window F1, F2, distant from the preceding revolution event by one revolution, that is to say by said large number N of teeth, and toleranced by +/ a tolerance n of teeth, that is to say having an extent of 2n teeth, is determined. An anomaly situated in such a window produces a revolution event. An anomaly situated outside such a window produces a no tooth event. The tolerance n is preferably equal to 2 teeth.

(52) In all the tests previously described, and mainly in the analysis step, where the term substantially equal is indicated, this expression means that the equality test is toleranced by +/ a tolerance p of teeth. The tolerance p is preferably equal to 2 teeth.