EP1497544B1 - Method for operating a fuel injection system for an internal combustion engine - Google Patents

Abstract

The invention relates to a method for operating a fuel injection system for an internal combustion engine during which it is monitored whether a time interval, in which a piezoelectric element for injecting fuel into a cylinder should be charged or discharged, overlaps a time interval, in which another piezoelectric element for injecting fuel into another cylinder should be charged or discharged. The invention is characterized in that it is monitored whether, during an injection of a lower priority, the charging or discharging occurs within a preset time interval around the moment of a charging or discharging of an injection of a higher priority. During the operation of the fuel injection system, the intervals of temporal charging and/or discharging flanks (flank overlappings) are determined, and the extent of the delay and/or shortening of the injections of a lower priority with regard to the injections of a higher priority is determined on the basis of this determination.

Description

State of the art

The invention relates to a method for operating a fuel injection system for an internal combustion engine according to the preambles of claims 1 and 6.

From DE 100 33 343 A1 a fuel injection system for an internal combustion engine, in particular a diesel engine is known, which has an injection control for monitoring and / or solving a conflict in driving the actuator elements, in particular a conflict management of superimposed injection curves of piezoelectric actuators.

With piezo common rail actuators, only one control edge can be executed at the same time. For reasons of combustion technology, however, it is necessary to apply the control of complementary banks in such a way that superimposed injections. This is then possible with the circuit arrangement known from DE 100 33 343 A1 for interconnecting piezoelectric elements, if the charge / discharge flanks of the piezoelectric elements have no overlap. In the case of overlapping flanks, it is provided in the case of the fuel injection system resulting from DE 100 33 343 A1 that the actuation is shifted or shortened with low priority (referred to hereinafter as low-priority actuation).

Further, Patent Abstracts of Japan Vol. 1999, No. 11, 30, 1999 (1999-09-03 & JP 11 159414 a (Hitachi Ltd), 15 June 1999 (1999-06-15) discloses a method of In this case, each injector is assigned an output stage consisting of a charging circuit and a discharge circuit In such an arrangement, the various injectors can be controlled simultaneously may be loaded and unloaded at the same time. This is prevented by a simple logic circuit that ensures that it is either charged or discharged.

It is the object of our invention to detect and determine edge overlaps, in particular between different injectors, and to deduce therefrom the necessary degree of time shift or shortening from the overlap area.

This object is achieved in a method for operating a fuel injection system of the type described above by the features of independent claim 1.

Advantageous embodiment of the method are the subject of the dependent claims.

Thus, the edge overlaps are advantageously determined during static and dynamic interrupts of a drive circuit during operation of the injection system. This determination preferably takes place as a function of the rotational speed and of the crankshaft angle of the internal combustion engine.

Single flank instants are examined in pairs for overlap.

Further advantages and details emerge from the following description of exemplary embodiments and the drawing.

Fig. 1

eine aus dem Stand der Technik bekannte Verschaltung piezoelektrischer Elemente;

Fig. 2a

das Laden eines piezoelektrischen Elementes;

Fig. 2b

das Laden eines piezoelektrischen Elementes;

Fig. 2c

das Entladen eines piezoelektrischen Elementes;

Fig. 2d

das Entladen eines piezoelektrischen Elementes;

Fig. 3

einen Ansteuerungs-IC;

Fig. 4

einen aus dem Stand der Technik bekannten zeitlichen Ablauf von Interrupts;

Fig. 5

schematisch eine Darstellung von Kollisionsbereichen von Flankenpaaren im Winkelbereich;

Fig. 6

eine schematische Darstellung des Verschiebens einer niederprioren Flanke nach spät und

Fig. 7

eine schematische Darstellung des Verkürzens einer niederprioren Ansteuerung.

In the drawing show:

Fig. 1

a known from the prior art interconnection piezoelectric elements;

Fig. 2a

the charging of a piezoelectric element;

Fig. 2b

the charging of a piezoelectric element;

Fig. 2c

the discharge of a piezoelectric element;

Fig. 2d

the discharge of a piezoelectric element;

Fig. 3

a drive IC;

Fig. 4

a timing of interrupts known in the art;

Fig. 5

schematically a representation of collision areas of edge pairs in the angular range;

Fig. 6

a schematic representation of the shift of a low-priority edge to late and

Fig. 7

a schematic representation of the shortening of a low-priority control.

Fig. 1 shows piezoelectric elements 10, 20, 30, 40, 50, 60 and means for their control. In this case, A denotes a region in a detailed representation and B a region in undetaillierter representation whose separation is indicated by a dashed line c. The detailed area A includes a circuit for charging and discharging the piezoelectric elements 10, 20, 30, 40, 50 and 60. In the example under consideration, the piezoelectric elements 10, 20, 30, 40, 50 and 60 are Actuators in fuel injection valves (in particular in so-called common rail injectors) of an internal combustion engine. In the described embodiment, six piezoelectric elements 10, 20, 30, 40, 50 and 60 are used to independently control six cylinders within an internal combustion engine; however, any other number of piezoelectric elements could be suitable for any other purposes.

The undetailed area B includes an injection control F with a controller D and a drive IC E serving to control the elements within the area A in detail. The driving IC E is supplied with various measured values of voltages and currents from the entire remaining driving circuit of the piezoelectric element. According to the invention, the control computer D and the drive IC E are designed to control the drive voltages and the drive times for the piezoelectric element. The control computer D and / or the drive IC E are also designed to monitor various voltages and currents of the entire drive circuit of the piezoelectric element.

In the following description, first of all, the individual elements are introduced within the area A shown in detail. The following is a general description of the processes of charging and discharging the piezoelectric elements 10, 20, 30, 40, 50 and 60. Finally, it will be described in detail how both processes are controlled and monitored by the control computer D and the driving IC E.

The piezoelectric elements 10, 20, 30, 40, 50 and 60 are divided into a first group G1 and a second group G2, each comprising three piezoelectric elements (ie, piezoelectric elements 10, 20 and 30 in the first group G1 and G1, respectively). piezoelectric elements 40, 50 and 60 in the second group G2). The groups G 1 and G 2 are components of parallel circuit parts. It can be determined with the group selection switches 310, 320 which of the groups G1, G2 of the piezoelectric elements 10, 20 and 30 or 40, 50 and 60 are each discharged by means of a common charging and discharging device (for charging processes, the group selection switches 310, 320 , as described in more detail below, but without meaning). The piezoelectric elements 10, 20 and 30 of the first group G 1 are arranged on an actuator bank and the piezoelectric elements 40, 50 and 60 in the second group G 2 are arranged on a further actuator bank. In this case, an actuator bank is a block in which two or more actuator elements, in particular piezoelectric elements, are permanently assigned, e.g. shed, are.

The group select switches 310, 320 are arranged between a coil 240 and the respective groups G1 and G2 (their coil-side terminals) and are realized as transistors. Drivers 311, 321 are implemented that transform control signals received from the drive IC E into voltages that are selectable as needed to close and open the switches.

Parallel to the group select switches 310, 320 are diodes 315 and 325, respectively (referred to as group select diodes). For example, when the group select switches 310, 320 are implemented as MOSFETs or IGBTs, these group select diodes 315 and 325 may be formed by the parasitic diodes themselves. During loads, the group select switches 310, 320 are bypassed by the diodes 315, 325. The functionality of the group selectors 310, 320 is therefore reduced to the selection of a Group G1, G2 of the piezoelectric elements 10, 20 and 30 or 40, 50 and 60 only for a discharge process.

Within the groups G1 and G2, respectively, the piezoelectric elements 10, 20 and 30 or 40, 50 and 60 are arranged as components of the parallel-connected piezo branches 110, 120 and 130 (group G1) and 140, 150 and 160 (group G2). Each piezoelectric branch comprises a series connection consisting of a first parallel connection with a piezoelectric element 10, 20, 30, 40, 50 and 60, and a resistor 13, 23, 33, 43, 53 or 63 (designated as branch resistor) and a second one Parallel connection with a designed as a transistor 11, 21, 31, 41, 51 and 61 (referred to as a selector switch) selector switch and (referred to as a branch diode) diode 12, 22, 32, 42, 52 and 62).

The branch resistors 13, 23, 33, 43, 53 and 63 cause the respective corresponding piezoelectric element 10, 20, 30, 40, 50 and 60 to discharge continuously during and after a charging process, since they each have both terminals of the capacitive connect piezoelectric elements 10, 20, 30, 40, 50 and 60 together. However, the branch resistors 13, 23, 33, 43, 53, and 63, respectively, are of sufficient size to make this process slow compared to the controlled charging and discharging operations, as described below. Therefore, the charge of any piezoelectric element 10, 20, 30, 40, 50, 60 within a relevant time after charging is considered to be fixed.

The branch selector switches / branch diode pairs in the individual piezo branches 110, 120, 130, 140, 150 or 160, ie, selector switch 11 and diode 12 in piezo branch 110, selector switch 21 and diode 22 in piezo branch 120, etc., can be realized as electronic switches (ie transistors) with parasitic diodes, for example MOSFETs or IGBTs (as indicated above for the group selector switch / diode pairs 310 and 315 or 320 and 325).

By means of the branch selector switches 11, 21, 31, 41, 51 and 61, it can be determined which of the piezoelectric elements 10, 20, 30, 40, 50 and 60 are each charged by means of a common charging and discharging device those Piezoelectric elements 10, 20, 30, 40, 50 and 60, respectively, whose branch selector switches 11, 21, 31, 41, 51 and 61 are closed during the charging process described below. Usually only one of the branch selector switches is closed.

The branch diodes 12, 22, 32, 42, 52 and 62 serve to bridge the branch selector switches 11, 21, 31, 41, 51 and 61 during discharging operations. Therefore, in the considered example of charging, each individual piezoelectric element can be selected, while for discharging, either the first group G1 or the second group G2 of the piezoelectric elements 10, 20 and 30 or 40, 50 and 60 or both must be selected ,

Returning to the piezoelectric elements 10, 20, 30, 40, 50 and 60 themselves, the branch selection piezo ports 15, 25, 35, 45, 55 and 65, respectively, can be operated either by means of the branch selection switches 11, 21, 31, 41, 51 and 61, respectively or via the corresponding diodes 12, 22, 32, 42, 52 and 62 and in both cases additionally via resistor 300 to ground.
By means of the resistor 300, the currents flowing during the charging and discharging of the piezoelectric elements 10, 20, 30, 40, 50 and 60 between the branch selection piezo-terminals 15, 25, 35, 45, 55 and 65, respectively, are measured. A knowledge of these currents allows a controlled charging and discharging of the piezoelectric elements 10, 20, 30, 40, 50 and 60. In particular, by closing and opening the charging switch 220 and discharge switch 230, depending on the magnitude of the currents, it is possible the charging current or set discharge current to predetermined average values and / or to prevent that they exceed or fall below predetermined maximum values and / or minimum values.

In the example considered, the measurement itself still requires a voltage source 621 which supplies a voltage of, for example, 5 V DC, and a voltage divider in the form of two resistors 622 and 623. Thus, the drive IC E (which carries out the measurements) should to be protected from negative voltages which might otherwise occur at measuring point 620 and which are not controllable by the driving IC E: Such negative voltages are added by addition to one of those mentioned Voltage source 621 and the voltage divider resistors 622 and 623 supplied positive voltage arrangement changed.

The other terminal of the respective piezoelectric element 10, 20, 30, 40, 50 and 60, i. the respective Gruppenwahlpiezoanschluß 14, 24, 34, 44, 54 and 64, via the group selector switch 310 or 320 or via the group selection diode 315 or 325 and a coil 240 and a parallel circuit consisting of a charging switch 220 and a charging diode 221 at the plus pole of a voltage source are connected, and alternatively or additionally via the group selector switch 310 or 320 or via the diode 315 or 325 and via the coil 240 and a parallel circuit consisting of a discharge switch 230 and a discharge diode 231 are grounded. Charging switch 220 and discharge switch 230 are realized, for example, as transistors, which are driven by drivers 222 and 232, respectively.

The voltage source comprises a capacitor 210. The capacitor 210 is charged by a battery 200 (for example, a motor vehicle battery) and a downstream DC-DC converter 201. The DC-DC converter 201 converts the battery voltage (for example, 12 V) into substantially any other DC voltage (for example, 250 V), and charges the capacitor 210 to this voltage. The control of the DC-DC converter 201 via the transistor switch 202 and the resistor 203, which serves the measurement of tapped at the measuring point 630 currents.

For the purpose of cross-checking is made possible by the driving IC E and by the resistors 651, 652 and 653 and, for example, a 5 V DC voltage source 654, a further current measurement at the measuring point 650; Furthermore, by the driving IC E and the voltage dividing resistors 641 and 642, a voltage measurement at the measuring point 640 is possible.

A resistor 330 (referred to as a total discharge resistor), a switch 331 (referred to as a stop switch) and a diode 332 (referred to as a total discharge diode) serve to discharge the piezoelectric elements 10, 20, 30, 40, 50 and 60 (if outside the normal mode , as described below, not by discharge the normal discharge operation). The stop switch 331 is preferably closed after □ normal □ discharging (cyclic discharging via discharging switch 230), thereby grounding the piezoelectric elements 10, 20, 30, 40, 50 and 60 through the resistors 330 and 300. Thus, any residual stresses remaining in the piezoelectric elements 10, 20, 30, 40, 50 and 60 are eliminated. The total discharge diode 332 prevents the occurrence of negative voltages on the piezoelectric elements 10, 20, 30, 40, 50 and 60, which could possibly be damaged by the negative voltages.

The charging and discharging of all the piezoelectric elements 10, 20, 30, 40, 50 and 60, or of a specific piezoelectric element 10, 20, 30, 40, 50 and 60, takes place with the aid of a single (all groups and their piezoelectric elements common) loading and unloading. In the example considered, the common charging and discharging device comprises the battery 200, the DC-DC converter 201, the capacitor 210, the charging switch 220 and the discharge switch 230, charging diode 221 and discharge diode 231, and the coil 240.

The charging and discharging of each piezoelectric element is performed in the same manner and will be explained below with reference to only the first piezoelectric element 10.

The states occurring during the charging and discharging operations will be explained with reference to FIGS. 2A to 2D, of which FIGS. 2A and 2B illustrate the charging of the piezoelectric element 10, and FIGS. 2C and 2D illustrate the discharging of the piezoelectric element 10.

The control of the selection of one or more to be loaded or unloaded piezoelectric elements 10, 20, 30, 40, 50 and 60, the charging process described below and the discharge takes place by the drive IC E and the control unit D by opening or Closing one or more of the switches 11, 21, 31, 41, 51, 61 introduced above; 310, 320; 220, 230 and 331. The interactions between the elements within the detailed area A shown on the one hand and the drive IC E and the control computer D on the other hand will be explained in more detail below.

With regard to the charging process, a piezoelectric element 10, 20, 30, 40, 50 or 60 to be charged must first be selected. To load only the first piezoelectric element 10, the branch selector switch 11 of the first branch 110 is closed, while all the remaining branch selector switches 21, 31, 41, 51, and 61 remain open. To load only any other piezoelectric element 20, 30, 40, 50, 60, or to load a plurality of simultaneously, its / their selection would be done by closing the corresponding branch selector switches 21, 31, 41, 51, and / or 61 ,

Then the charging process itself can take place:

Within the example considered, a positive potential difference between the capacitor 210 and the group select piezo terminal 14 of the first piezoelectric element 10 is generally required for the charging process. However, as long as charging switch 220 and discharging switch 230 are opened, no charging or discharging of the piezoelectric element 10 takes place. In this state, the circuit shown in Fig. 1 is in a stationary state, i. the piezoelectric element 10 maintains its charge state substantially unchanged, with no currents flowing.

To charge the first piezoelectric element 10, switch 220 is closed. Theoretically, the first piezoelectric element 10 could be charged by it alone. However, this would lead to large currents that could damage the elements concerned. Therefore, the currents occurring at the measuring point 620 are measured and switch 220 is opened again as soon as the detected currents exceed a certain limit. Thus, to achieve any charge on the first piezoelectric element 10, charging switch 220 is repeatedly closed and opened while discharge switch 230 remains open.

Upon closer inspection, the charging conditions shown in FIG. 2A result when the charging switch 220 is closed, ie, a closed circuit comprising a series circuit consisting of the piezoelectric element 10, capacitor 210 and the coil 240, wherein in the circuit, a current i _LE (t) flows, as indicated in Fig. 2A by arrows. Due to this current flow, both the group selection piezo terminal 14 of the first piezoelectric element 10 is supplied with positive charges and energy stored in the coil 240.

When the charging switch 220 opens shortly (for example a few μs) after closing, the conditions shown in FIG. 2B result: a closed circuit comprising a series arrangement consisting of the piezoelectric element 10, discharge diode 231 and coil 240, wherein in the circuit a current i _LA (t) flows, as indicated in Fig. 2B by arrows. Due to this current flow, energy stored in the coil 240 flows into the piezoelectric element 10. According to the power supply to the piezoelectric element 10, the voltage occurring therein increases and increases its outer dimensions. When the energy has been transferred from the coil 240 to the piezoelectric element 10, the stationary state of the circuit shown in FIG. 1 and already described is reached again.

At this time, or sooner or later (depending on the desired time profile of the charging process), charging switch 220 is again closed and reopened so that the operations described above occur again. Due to the re-closing and reopening of the charging switch 220, the energy stored in the piezoelectric element 10 increases (the energy already stored in the piezoelectric element 10 and the newly added energy add up), and the voltage appearing on the piezoelectric element 10 increases and its outer dimensions increase accordingly.

If the above-mentioned closing and opening of the charging switch 220 are repeated many times, the increase of the voltage appearing on the piezoelectric element 10 and the extension of the piezoelectric element 10 are stepwise.

When the charging switch 220 has been closed and opened a predetermined number of times and / or the piezoelectric element 10 has reached the desired state of charge, the charging of the piezoelectric element is stopped by leaving the charging switch 220 open.

With respect to the unloading operation, in the example under consideration, the piezoelectric elements 10, 20, 30, 40, 50 and 60 are discharged into groups (G 1 and / or G 2) as described below:

First, the group selector switch 310 and / or 320 of the group G1 and / or G2 whose piezoelectric elements are to be discharged are closed (the selector switches 11, 21, 31, 41, 51, 61 have no influence on the selection of the piezoelectric elements 10, 20, 30, 40, 50, 60 for the unloading operation, since in this case they are bridged by the diodes 12, 22, 32, 42, 52 and 62). Therefore, in order to discharge the piezoelectric element 10 as part of the first group G1, the first group selector switch 310 is closed.

When the discharge switch 230 is closed, the conditions shown in FIG. 2C result: a closed circuit is formed comprising a series arrangement consisting of the piezoelectric element 10 and the coil 240, wherein a current i _EE (t) flows in the circuit, such as in Fig. 2C indicated by arrows. Due to this current flow, the energy stored in the piezoelectric element (a part thereof) is transferred to the coil 240. According to the energy transfer from the piezoelectric element 10 to the coil 240, the voltage appearing on the piezoelectric element 10 decreases and decreases in its outer dimensions.

When the discharge switch 230 opens shortly (for example, a few μs) after closing, the conditions shown in FIG. 2D result: a closed circuit comprising a series arrangement consisting of the piezoelectric element 10, capacitor 210, charging diode 221 and the coil 240 in which a current i _EA (t) flows in the circuit, as indicated by arrows in FIG. 2D. Due to this current flow, energy stored in the coil 240 is returned to the capacitor 210. When the energy has been transferred from the coil 240 to the capacitor 210, the stationary state of the circuit shown in FIG. 1 and already described is reached again.

At this time, or sooner or later (depending on the desired time profile of the discharge process), discharge switch 230 is again closed and reopened so that the proceeding again described above. Due to the re-closing and reopening of the discharge switch 230, the energy stored in the piezoelectric element 10 further decreases, and the voltage appearing on the piezoelectric element and its external dimensions also decrease accordingly.

When the above-mentioned closing and opening of the discharge switch 230 are repeated many times, the decrease of the voltage appearing on the piezoelectric element 10 and the extension of the piezoelectric element 10 are stepwise.

When discharging switch 230 has been closed and opened a predetermined number of times and / or the piezoelectric element has reached the desired state of charge, discharging of the piezoelectric element is stopped by leaving the discharging switch 230 open.

The interaction between the drive IC E and the control computer D on the one hand and the elements within the detailed area A on the other hand by means of control signals via branch selection control lines 410, 420, 430, 440, 450, 460, group selection control lines 510, 520, stop switch control line 530, charge switch control line 540 and discharge switch control line 550, and control line 560 are supplied to elements within the detailed area A from the drive IC E. On the other hand, at the measuring points 600, 610, 620, 630, 640, 650 within the range A shown in detail, sensor signals are detected which are supplied to the driving IC E via the sensor lines 700, 710, 720, 730, 740, 750.

For selecting the piezoelectric elements 10, 20, 30, 40, 50 and 60, respectively, for carrying out charging or discharging operations of individual or several piezoelectric elements 10, 20, 30, 40, 50, 60 by opening and closing the corresponding switches, such as As described above, voltages are applied to the transistor bases by means of the control lines. By means of the sensor signals, in particular, a determination of the resulting voltage of the piezoelectric elements 10, 20 and 30, or 40, 50 and 60 based on the measuring points 600 and 610 and the charging and discharging by means of the measuring point 620.

In Fig. 3, some of the components contained in the drive IC E are indicated: a logic circuit 800, memory 810, digital-to-analog converter 820, and comparator module 830. It is further noted that the fast parallel bus 840 (used for control signals) is connected to the logic circuit 800 of the drive IC E, while the slower serial bus 850 is connected to the memory 810. The logic circuit 800 is connected to the memory 810, to the comparator module 830 as well as to the signal lines 410, 420, 430, 440, 450 and 460; 510 and 520; 530, 540, 550 and 560 connected. The memory 810 is connected to the logic circuit 800 and to the digital-to-analog converter module 820. Furthermore, the digital-to-analog converter module 820 is connected to the comparator module 830. In addition, the comparator chip 830 is connected to the sensor lines 700 and 710, 720, 730, 740 and 750 and - as already mentioned - to the logic circuit 800.

4 schematically shows a time sequence of interrupts known from the prior art for programming the beginning of a main injection HE to be described in more detail below, as well as two pilot injections VE1 and VE2 as a function of top dead center of the crankshaft. As can be seen from FIG. 4, in a 6-cylinder engine, static interrupts are generated at, for example, about 78.degree. Crankshaft and, for example, about 138.degree. Crankshaft, by which the beginning of pre-injection VE2 as well as directly before main injection HE preinjection VE1 be programmed. The ends of these injections are then programmed based on dynamic interrupts. It is understood that the above crankshaft angles are merely exemplary indications. In principle, the interrupts can also be generated at other crankshaft angles. Above, only the programming of pilot injections has been explained. In a corresponding manner, however, to proceed with post-injection, if such are made.

The calculation for edge overlap detection occurs in every static and dynamic interrupt. Only overlaps between edges that are known at the time of the interrupts can be calculated.

• 1. Die aktuelle Drehzahl n wird ermittelt, diese Drehzahl n wird im gesamten Interrupt verwendet ("Einfrieren der Drehzahl");
• 2. mit jedem Interrupt werden neue Informationen über Flanken bekannt. Damit nur aktuelle Informationen paarweise verglichen werden, wird der Informationsstatus aktualisiert. Bei jedem Interrupt wird daher ein Flag für neue Informationen gesetzt und geprüft, ob Ansteuerungen, bei denen Flags gesetzt sind, bereits abgearbeitet sind. In diesem Fall werden die betreffenden Flags gelöscht;
• 3. es folgt eine Bestimmung der Zeitpunkte der Flankenbearbeitung bezogen auf eine beliebige Referenz, zum Beispiel auf die Referenzzeit t = 0 bei einem Kurbelwellenwinkel phi = 0°. Dabei werden die bekannten Informationen über Beginnwinkel, Zeitoffset, Beginn und Dauer unter Berücksichtigung der aktuellen Drehzahl zur Extrapolation benutzt. Der allgemeine Zusammenhang zwischen Drehzahl n, Winkel phi und Zeit t ist: $$\mathrm{n}=(\mathrm{phi}/\mathrm{t})*\mathrm{c}=\mathrm{Gleichung}$$
  
  
  wobei die Zeit in µ-Sekunden, der Kurbelwellenwinkel phi in °KW gemessen werden und die Konstante c 166667 (U/min) / (°KW/µs) beträgt;
• 4. die einzelnen Flankenzeitpunkte werden paarweise auf Überlappung untersucht. Vorteilhafterweise werden nur Paare gemischter Bankzugehörigkeit getestet, da sich Überlappungen innerhalb derselben Bank aus Applikationsfehlern ergeben. Die sichere Strategie besteht aber dennoch im Testen jeder denkbaren Flankenpaare;
• 5. jeder Einspritzung wird eine Priorität zugeteilt. Abhängig von System- und Umgebungsparametern wird jeder Einspritzung eine bestimmte Priorität zugeordnet. Dadurch wird bei jeder Einspritzpaarung unterschieden zwischen niederprioren und hochprioren Ansteuerungen. Es wird sichergestellt, daß eine Umschaltung der Prioritäten während eines Berechnungsablaufs keine negativen Auswirkungen hat. So können beispielsweise nach der aktuellen Prioritätenkonstellation eine Überlappungserkennung und Maßnahmen im statischen Interrupt vorgenommen werden, anschließend können die Prioritäten umgeschaltet werden, also geändert werden. In den nachfolgenden dynamischen Interrupts dieser Paarung müßte nach neuer Priorität regiert werden, was im schlimmsten Falle eine Maßnahme gegen die Ansteuerung einer Einspritzung höherer Priorität (hochpriore Ansteuerung) zur Folge hätte.
  Deshalb muß auch bei einer derartigen Umschaltung der Prioritäten die Konsistenz der Prioritätenzuordnung gewährleistet sein. Vorteilhaft wird dies durch Zuordnung eines Prioritätensatzes zu jeder Paarung erfüllt. Die Größe des Buffers für verschiedene Prioritätensätze muß dabei so gewählt werden, daß die maximal mögliche Anzahl an Änderungen der Prioritäissätze während der gesamten Abarbeitung einer Paarung gespeichert werden kann. Der Prioritätensatz einer Paarung wird nach ihrer vollständigen Abarbeitung mit dem aktuellen, durch einen Prioritätenmanager der elektronischen Ansteuerschaltung vorgegebenen Satz erneuert;
• 6. bei der Untersuchung auf Kollision wird in der Zeitbasis der Abstand des jeweiligen Beginns der beiden Flanken zueinander ermittelt. Ausgehend von diesem Abstand kann entschieden werden, ob eine Überlappung vorliegt. Da die Flankenzeitpunkte auf den Winkeln der Einspritzungen aufbauen, muß hierbei auf 720°KW-Überläufe besonders geachtet werden. Hierbei ist rein prinzipiell eine Vielzahl von Realisierungsmöglichkeiten bei der Abstandsberechnung und der Auswertung denkbar. In der nachfolgend beschriebenen Ausgestaltung des Verfahrens werden 3 Berechnungen vorgenommen.
  In Fig. 5 sind die Berechnungen auf der Winkelbasis dargestellt, auf der Abszisse ist dabei der Wert einer niederprioren Flanke A, auf der Ordinate der Wert einer hochprioren Flanke B aufgetragen. Die hochpriore Flanke wird nach früh (pre) und nach spät (post) mit Bereichen "abgesichert". Falls nun eine niederpriore Flanke diesen Bereich schneidet, liegt eine Kollision vor. Die Bereiche sind in der Abbildung markiert. Bereiche außerhalb 720°KW = phi_max werden entsprechend der Zuordnung in den zulässigen Bereichen übertragen. Die Ergebnisse der Berechnungen gemäß folgender Gleichungen: $$\mathrm{B}-\mathrm{A}=\mathrm{Gleichung}$$
  
  $$\mathrm{B}-\mathrm{A}-{\mathrm{phi}}_{\max }=\mathrm{Gleichung}$$
  
  $$\mathrm{B}-\mathrm{A}+{\mathrm{phi}}_{\max }=\mathrm{Gleichung}$$
  
  sind im Diagramm in Fig. 2 gekennzeichnet. Überlappungen, die durch die einzelnen Berechnungen erkannt werden, sind dabei jeweils durch die gleiche Schraffur gekennzeichnet. In Fig. 5 ist der Zusammenhang auf Winkelbasis erläutert, die Transformation in den Zeitbereich erfolgt mit der oben erläuterten Gleichung (1). Ein Beispiel von A = 50° und B = 100° liefert mit Gleichung (2) die Überlappung bei gegebenen Werten der Verschiebung nach früh (pre) und nach spät (post);
• 7. In Abhängigkeit vom Überlappungsgrad wird der Grad der Verschiebung bzw. Verkürzung ermittelt. Verschoben wird nach spät derart, daß die niederpriore Beginnflanke im Abstand eines Zeitvorhalts nach dem voraussichtlichen Ende der hochprioren Flanke plaziert wird. Beim Verschieben wird die Dauer beibehalten. Verschoben wird auch der Zeitpunkt des dynamischen Interrupts, der an die Beginnflanke mit festem Abstand gekoppelt ist. Verkürzt wird derart, daß die niederpriore Endeflanke nach früh verschoben wird. Der Zeitpunkt der Beginnflanke wird beibehalten. Die Entscheidung ob verschoben oder verkürzt wird, hängt davon ab, ob zum Zeitpunkt der Überlappungserkennung die Beginnflanke bereits abgearbeitet wird. Wird die Beginnflanke bereits abgearbeitet, wobei hierunter der Beginn des Ablaufs des Verbrennungsvorgangs verstanden wird, so ist ein Verschieben nicht mehr möglich, es kann nur noch verkürzt werden. Daraus folgt, daß bei allen Überlappungen von niederprioren Endeflanken nur verkürzt werden kann, da der Zeitpunkt der Überlappungserkennung nur im dynamischen Interrupt der niederprioren Einspritzung liegen kann, diese aber mit der Ausfiihrung der Beginnflanke verbunden ist.

In each interrupt, the following steps are performed:

• 1. The current speed n is determined, this speed n is used throughout the interrupt ("freezing the speed");
• 2. With each interrupt, new information about edges is known. So that only current information is compared in pairs, the information status is updated. For each interrupt, therefore, a flag for new information is set and checked as to whether actuations in which flags are set have already been executed. In this case, the relevant flags are deleted;
• 3. It follows a determination of the timing of the edge processing with respect to any reference, for example, the reference time t = 0 at a crankshaft angle phi = 0 °. The known information about the start angle, time offset, start and duration are used, taking into account the current speed for extrapolation. The general relationship between speed n, angle phi and time t is: $$\mathrm{n}=(\mathrm{phi}/\mathrm{t})*\mathrm{c}=\mathrm{equation}$$
  
  
  where the time in μ seconds, the crankshaft angle phi are measured in ° CA and the constant c is 166667 (rpm) / (° CA / μs);
• 4. The individual flank times are examined in pairs for overlap. Advantageously, only pairs of mixed bank affiliation are tested because overlaps within the same bank result from application errors. The safe strategy, however, is to test every imaginable flank pair;
• 5. Each injection is assigned a priority. Depending on system and environmental parameters, each injection is assigned a specific priority. This distinguishes between low-priority and high-priority actuations for each injection pairing. It ensures that a change of priorities during a calculation process has no negative effects. Thus, for example, according to the current priority constellation, an overlap detection and measures in the static interrupt can be made, then the priorities can be switched over, that is to say changed. In the subsequent dynamic interrupts of this pairing would have to be governed by new priority, which in the worst case would have a measure against the control of a higher priority injection (high-priority control) would result.
  Therefore, even with such a switchover of priorities, the consistency of the priority allocation must be guaranteed. This is advantageously accomplished by assigning a priority set to each pairing. The size of the buffer for different priority sets must be selected so that the maximum possible number of changes in the priority rates can be stored during the entire processing of a pairing. The priority set of a pairing is renewed after its complete execution with the current set given by a priority manager of the electronic control circuit;
• 6. In the investigation of a collision, the distance of the respective beginning of the two flanks from one another is determined in the time base. Based on this distance, it can be decided whether an overlap exists. Since the flank time points build on the angles of the injections, particular care must be taken here for 720 ° CA overflows. Here, in principle, a variety of implementation options in the distance calculation and the evaluation is conceivable. In the embodiment of the method described below, 3 calculations are made.
  In FIG. 5, the calculations are shown on the angle basis, on the abscissa is the value of a low-priority edge A, plotted on the ordinate, the value of a high-priority edge B. The high-priority edge is "secured" early (pre) and late (post) with areas. If a low-slope edge intersects this area, there is a collision. The areas are marked in the picture. Areas outside 720 ° KW = phi _max are transmitted according to the assignment in the permissible ranges. The results of the calculations according to the following equations: $$\mathrm{B}-\mathrm{A}=\mathrm{equation}$$
  
  $$\mathrm{B}-\mathrm{A}-{\mathrm{phi}}_{\mathrm{Max}}=\mathrm{equation}$$
  
  $$\mathrm{B}-\mathrm{A}+{\mathrm{phi}}_{\mathrm{Max}}=\mathrm{equation}$$
  
  are marked in the diagram in FIG. Overlaps that are detected by the individual calculations are each characterized by the same hatching. In Fig. 5, the relationship is explained on an angle basis, the transformation into the time domain is carried out with the above-explained equation (1). An example of A = 50 ° and B = 100 ° yields with equation (2) the overlap at given values of the shift to early (pre) and to late (post);
• 7. Depending on the degree of overlap, the degree of displacement or shortening is determined. Shifting is delayed so that the low-priority starting edge is placed at a distance of a time lead after the expected end of the high-priority edge. When moving, the duration is retained. The time of the dynamic interrupt, which is coupled to the start edge with a fixed distance, is also shifted. Is shortened such that the low-priced end flank is moved to early. The timing of the beginning edge is maintained. The decision whether to postpone or shorten depends on whether the beginning edge is already processed at the time of overlap detection. If the starting edge has already been processed, this being understood as the beginning of the course of the combustion process, it is no longer possible to move it, it can only be shortened. It follows that in all overlaps of low-priority end flanks can only be shortened because the time of overlap detection can only be in the dynamic interrupt of the lower priority injection, but this is associated with the execution of the beginning edge.

As an example, a shift is shown in connection with FIG. The overlap is recognized by equation (2), the resulting overlap amount _tk is directly related to the degree of shift. The degree of shift is $$\begin{array}{l}{\mathrm{t}}_{\mathrm{k}}+\mathrm{time\; allowance}+\mathrm{Hedging\; area\; post}\\ =\mathrm{equation}\end{array}$$

Equation (5) holds even if the overlap was determined from Equation (3) or Equation (4).

For a shortening of the drive time, an example is shown in FIG. The overlap is again recognized by equation (2), the resulting overlap amount t _k is here also directly in the degree of shortening. The degree of shortening is $$\begin{array}{l}{\mathrm{t}}_{\mathrm{k}}+\mathrm{time\; allowance}+\mathrm{Hedging\; area\; pre}\\ =\mathrm{equation}\end{array}$$

Equation (6) holds even if the overlap was determined from equation (3) or equation (4).

In addition to primary collisions, secondary collisions are also possible. Secondary collisions occur, for example, when the low-early beginning edge is moved late in the static interrupt, but this collides with the high-priority end edge. The time of the collision detection is then in the dynamic interrupt of the high-priority control. Thus, the low-pried start edge must be moved further late in this secondary collision. The same applies in the case of tertiary collisions. An advantageous embodiment of the method provides that after a review of all pairings that ended with the detection of an overlap and associated action, a repeated cycle of all pairings takes place, namely until either a termination criterion based on the number of passes occurs or overlap is detected.

In another embodiment of the method, the detection of undesired overlaps of the time intervals in which a piezoelectric element is to be charged or discharged with a time interval in which the other piezoelectric Element is to be loaded or unloaded by calculating used angle ranges and comparing with predetermined allowable angle ranges, that is, collision-free or collision-tolerant angle ranges.

A collision-free angular range is understood to mean the angular range which may be swept by the various injection types of a cylinder of the internal combustion engine without overlapping of activations of the actuators. The collision-free angle range is determined, for example, in a 4-cylinder internal combustion engine with a 1-bank structure by dividing the value 720 ° crankshaft angle by the number of cylinders, ie four. The collision-free angle range is thus 180 ° crankshaft angle in an internal combustion engine of this type. The used angular range is the crankshaft angle range swept from the beginning of the earliest pilot injection to the end of the latest post-injection. If the angular range used now exceeds the collision-free angular range, for example, a late injection of one cylinder overlaps with an early injection of another cylinder on the same bank. As already mentioned above, only one actuator may be loaded at the same time on one bank, otherwise a charge compensation would take place which can lead to a disturbed activation.

In addition to the 1-bank structure, a plurality of cylinders can be combined to form a bank, with several banks being driven by the same supply unit for loading or unloading. Such an arrangement is called a quasi-multi-bank structure. In this case, the angular range in which collisions of drives on different banks can be resolved by edge management is called a collision tolerant range. In this case, exceeding the collision-tolerant plus collision-free angle range leads to disturbed drive characteristics.

Using the example of a 6-cylinder internal combustion engine with a quasi-2-bank structure, the collision-free angle range is 120 ° crankshaft angle and the collision-tolerant angular range is also 120 ° crankshaft angle. The total allowable angle range is now the sum of the collision-free angle range and the collision-tolerant angular range, in the case of the 6-cylinder quasi-2-bank internal combustion engine, the allowable angular range is 240 ° crankshaft angle. Generally, in a quasi-2-bank internal combustion engine, the allowable angular range can be determined by dividing the value of 720 ° crank angle by the number of cylinders multiplied by the number of banks.

The core of this embodiment of the method for operating a fuel injection system for an internal combustion engine is the calculation of the used angular range and the comparison with the permissible angular range, that is the collision-free or the sum of collision-free and collision-tolerant angular range.

Hereinafter, an embodiment of the method will be described.

• 1. Die aktuelle Drehzahl n wird ermittelt, diese Drehzahl wird im gesamten Interrupt verwendet ("Einfrieren der Drehzahl").
• 2. Mit jedem Interrupt werden neue Informationen über Flanken bekannt. Diese Informationen werden unter Verwendung der aktuellen Drehzahl n auf die Winkelbasis umgerechnet.
• 3. Jede neu hinzugekommene Winkelinformation wird in den genutzten Winkelbereich hineingerechnet. Dabei wird aus der Menge der bekannten Winkelinformationen eine Minimum-/Maximumauswahl vorgenommen mit dem Ziel, die zu einem Arbeitsspiel gehörende früheste und späteste Ansteuerflanke zu ermitteln. Aus den Winkelinformationen der frühesten und spätesten Ansteuerflanken wird durch Differenzbildung der bekannte genutzte Winkelbereich ermittelt.
  Nach dem dynamischen Interrupt der letzten Nacheinspritzung ist auf diese Weise der gesamte genutzte Winkelbereich von der frühesten Voreinspritzung bis zur spätesten Nacheinspritzung bekannt, wobei der allgemeine Zusammenhang zwischen Drehzahl n Winkel phi und Zeit t oben in Form der Gleichung (1) bereits erläutert wurde.
• 4. Der bekannte genutzte Winkelbereich wird mit den vorgegebenen kollisionsfreien und kollisionstoleranten Winkelbereichen verglichen. Bei Bereichsüberschreitung erfolgt eine Fehlermeldung und eine Quantifizierung der Bereichsüberschreitung.
• 5. Bei allen Berechnungen wird dabei die Drehzahldynatnik mit ihrer Wirkung vom Berechnungszeitpunkt bis zum Zeitpunkt der Abarbeitung, das heißt der Ansteuerung der Aktoren berücksichtigt.

In each interrupt, new information used to calculate the angular range used is known. The following steps are performed in each interrupt:

• 1. The current speed n is determined, this speed is used throughout the interrupt ("freezing the speed").
• 2. With each interrupt, new information about edges is known. This information is converted to the angle base using the current speed n.
• 3. Each newly added angle information is calculated into the used angle range. In this case, a minimum / maximum selection is made from the set of known angle information with the aim of determining the earliest and latest control edge belonging to a working cycle. From the angle information of the earliest and latest control edges, the known angular range used is determined by subtraction.
  After the dynamic interrupt of the last post-injection in this way, the entire used angle range from the earliest pilot injection to the latest post-injection is known, the general relationship between speed n angle phi and time t above in the form of equation (1) has already been explained.
• 4. The known used angular range is compared with the predetermined collision-free and collision-tolerant angular ranges. If the range is exceeded, an error message is issued and the range is exceeded.
• 5. For all calculations, the speed dynatnik is taken into account with its effect from the calculation time to the time of execution, that is to say the actuation of the actuators.

• a) eine entsprechende Verschiebung einer niederprioren Einspritzung, so daß der genutzte Winkelbereich wieder im zulässigen Bereich liegt;
• b) eine Berücksichtigung der Fehlermeldung und des Grades der Bereichsüberschreitung bei der nächstfolgenden Ansteuerung im gleichen oder ähnlichen Betriebspunkt.

Ways to respond to an error message are now

• a) a corresponding shift of a low-priority injection, so that the used angle range is again in the allowable range;
• b) taking into account the error message and the degree of overshoot at the next activation in the same or similar operating point.

Claims (5)

1. Method for operating a fuel injection system for an internal combustion engine having at least two cylinders which are each assigned to a bank, the fuel injection system having at least two piezo-electric elements and each cylinder being associated with in each case at least one piezo-electric element for injecting fuel into the cylinder by charging or discharging the piezo-electric element, and wherein the piezo-electric elements are associated with a supply unit for charging or discharging the piezo-electric element, and wherein monitoring is carried out to determine whether an overlap occurs between a time interval in which a piezo-electric element is to be charged or discharged and a time interval in which the other piezo-electric element is to be charged or discharged, characterized in that monitoring is carried out to determine whether, when there is a relatively low priority injection, the charging or discharging takes place within a predefined time interval around the time of charging or discharging for a relatively high priority injection, wherein the distances between charging and/or discharging time signal edges are determined while the fuel injection system is operating and the magnitude of the displacement and/or shortening of the relatively low priority injections compared to the relatively high priority injections are determined therefrom.
2. Method according to Claim 1, characterized in that the priorities of the injection are predefined, with the predefined value being retained for one injection cycle.
3. Method according to Claim 1 or 2, characterized in that the signal edge overlaps are determined during interrupts in a drive circuit while the fuel injection system is operating.
4. Method according to Claim 3, characterized in that the signal edge overlaps are determined as a function of the rotational frequency and the crankshaft angle of the internal combustion engine.
5. Method according to one of Claims 1 to 4, characterized in that the signal edge overlaps are determined in pairs, preferably for pairs with mixed bank association.

Images