DE102006044771B4 - Method and control unit for determining an error of an injection quantity of an injection control element of an internal combustion engine which is controlled with a control duration - Google Patents

Abstract

Method for determining the deviation (F) of the actual injection quantity from a predetermined reference quantity of a fuel injection system (12) of an internal combustion engine (10) having at least one controllable injection actuator (20) in which the injection quantity required at a specific operating point of the internal combustion engine (10) per combustion event (QLL) is injected for various combustion processes with different injection patterns, and wherein the deviation (F) is determined from a comparison of the reference quantity (QLx / x) formed in a first injection pattern with the reference quantity (QLy / y) formed in a second injection pattern characterized in that the first injection pattern (35) comprises x single injections and that the second injection pattern (37) comprises y single injections, where x and y are each greater than one, and x is not equal to y.

Description

State of the art

The invention relates to a method according to the preamble of claim 1 and a control device according to the preamble of the independent device claim. Such a method and such a control device is in each case from DE 103 43 759 A1 known.

Injection systems of modern internal combustion engines must meet high requirements for the precision of the injection times and injection quantities for optimum noise and exhaust gas performance of internal combustion engines. When controlling an injection actuator, for example an injector, due to production-related and aging-related tolerances, deviations of an actually injected injection quantity from the injection quantity calculated in the control unit of the internal combustion engine occur. These deviations can lead to a deterioration of the exhaust behavior, to noise problems, a running noise and unacceptable performance variations.

To determine the deviations sees the DE 103 43 759 A1 prior to metering the fuel required to maintain a constant operating point first with a single injection and then with a multiple injection of the same injection quantity. Since the error in the single injection occurs once and in the multiple injection several times, dependencies arise that allow a quantitative determination of the error. The knowledge of the error allows a compensation in principle.

Especially with small injection quantities, compensation of quantity errors is of great importance. Small injection quantities are metered, for example, in diesel engines in the form of pilot injections. The fulfillment of current and future noise and exhaust gas requirements places high demands on the temporal stability of the injected pre-injection quantity. Even minor changes in the actual injected pilot injection quantity can have a significant effect on the exhaust gas quality and the engine noise.

Disclosure of the invention

Against this background, the object of the invention is to specify a method and a control device with which the errors mentioned can be corrected more reliably, in particular with small injection quantities, so that the fluctuation range of remaining residual errors is lower than in the subject matter DE 103 43 759 A1 , This object is achieved in a method of the type mentioned by the characterizing features of claim 1 and in a control device of the type mentioned by the characterizing features of the independent device claim.

A major difference to the subject of the DE 103 43 759 A1 is that in the present invention for determining the deviation, two injection patterns are used, which have both multiple injections. The number of multiple injections differs by at least one injection. In the DE 103 43 759 A1 On the other hand, a single injection is correlated as a reference with a multiple injection.

In the known method, therefore, there is a comparatively large difference between the comparatively short drive times of the partial injections of the multiple injection patterns and the comparatively long drive duration of the single injection. Because of the comparatively large difference, it may happen that the error occurring in the comparatively short activation periods of the partial injections has a different value than the error occurring in the case of the large activation duration of the individual injection.

From the DE 60 2004 003 390 T2 For example, a method for real-time determination of a flow rate characteristic of a fuel injection nozzle is known.

The DE 10 2005 051 701 A1 describes a method for adapting a characteristic of an injection valve. The total injection amount is divided into a basic injection and at least one measurement injection. The injection quantity of the metering injection is successively reduced and the injection duration of the base injection is increased so that the total injection amount remains the same.

The DE 10 2006 019 894 B3 describes a method of operating an internal combustion engine having a plurality of cylinders. The cylinders are operated with metering pulses with different pulse characteristics.

The inventors have recognized that such deviations are responsible for residual defects that occur in particular with small injection quantities.

By contrast, in the invention using two multiple injection patterns to determine the error, much smaller distances occur between the drive times of the partial injections of the multiple injection patterns. This has the consequence that even the errors that occur at the aforementioned drive times, at best, may differ slightly. The result is a much more accurate determination of the error allows small injection quantities. This is a significant advantage especially in diesel engines, in which the injection quantity is divided in normal operation to at least a small pilot injection quantity and a larger main injection quantity.

The invention thus allows, in particular, a reliable detection of production-related variations of pilot injection quantities as well as reliable detection of a drift of a pilot injection quantity over the lifetime of an injection system. With the reliable detection also results in the possibility of reliable and accurate compensation of deviations. Thus, an undesirable failure of the pilot injection is inherently prevented because of a too short drive duration. The detection can be done using existing functions such as idle and smooth running control methods, so that the invention requires no additional sensors.

Further advantages will become apparent from the description and the accompanying figures.

list of figures

• 1 einen Verbrennungsmotor mit einem Steuergerät und einem Einspritzsystem;
• 2 ein Flussdiagramm, das die Bildung von Ansteuersignalen für das Einspritzsystem verdeutlicht;
• 3 Kennlinien der Abhängigkeit von Einspritzmengen von Ansteuerdauern zur Verdeutlichung eines bekannten Verfahrens;
• 4 vergleichbare Kennlinien zur Verdeutlichung eines Ausführungsbeispiels eines erfindungsgemäßen Verfahrens; und
• 5 ein Ausführungsbeispiel eines erfindungsgemäßen Verfahrens.

Embodiments of the invention are illustrated in the drawings and are explained in more detail in the following description. In each case, in schematic form:

• 1 an internal combustion engine with a control unit and an injection system;
• 2 a flowchart illustrating the formation of drive signals for the injection system;
• 3 Characteristics of the dependence of injection quantities of driving periods to illustrate a known method;
• 4 comparable characteristics to illustrate an embodiment of a method according to the invention; and
• 5 an embodiment of a method according to the invention.

Embodiment (s) of the invention

In detail, the shows 1 an internal combustion engine 10 with an injection system 12 , an angle sensor 14 , a control unit 16 and a driver's desire 18 , The injection system 12 has at least one injection actuator 20 on. The injection actuator 20 In one embodiment, an injector, with the fuel individually for a combustion chamber of the internal combustion engine 10 is measured. The internal combustion engine 10 is preferably a diesel engine, but the invention is not limited to diesel engines, but can also be implemented with gasoline engines and Wankel engines.

The driver's desire 18 serves to detect a torque request to the internal combustion engine 10 , Will the internal combustion engine 10 used as a drive motor of a road vehicle, the torque request is made by the driver. Alternatively or additionally, however, the torque request can also be predetermined by functions of the motor vehicle that are in the control unit 16 or other control devices of the motor vehicle or internal combustion engine 10 be performed. An example of one in the control unit 16 performed function is an idle speed control, which also without pressing the driver request generator 18 is active and the one idle speed of the engine 10 keeps constant. Depending on the torque requirements mentioned forms the control unit 16 including drive times AD with which the at least one injection actuator 20 is controlled. The driving time AD , and that during the drive AD via the injection actuator 20 metered amount of fuel Q represents a significant factor for determining the torque of the internal combustion engine 10 represents.

The angle sensor 14 is implemented in one embodiment as an inductive sensor, which scans ferromagnetic markings of a sensor wheel and thereby provides a rotational angle information ° KW. The control unit 16 forms from the rotation angle information ° KW in particular a value for the average speed of the internal combustion engine 10 as input value of the idle speed control and possibly even values of cylinder-individual deviations from the value of the average speed as an input variable of a known cylinder-individual quantity compensation control or running restraint. Furthermore, the rotation angle information ° KW allows the control unit 16 an output of the control signals AD at the correct position of pistons of the internal combustion engine 10 ,

Incidentally, the controller 16 to set up, in particular programmed to control the flow of one of the methods presented here and in particular control signals AD for controlling the injection actuator 20 to build.

2 shows a flowchart of an embodiment for the formation of actuating signals AD for the injection actuator 20 , The step represents 22 a major parent program HP for controlling the internal combustion engine 10 , From the main program 22 out becomes synchronous to the movement of pistons of the internal combustion engine 10 a step 24 achieved in which an underlying Qb one in a certain operating point of the internal combustion engine 10 amount of fuel to be injected Q is formed. The underlying Qb For example, it depends on the signal from the driver's request generator 18 or depending on the momentary intervention of a Vehicle dynamics control ( ESP ) or depending on pilot control variables of an idle controller and / or a Laufruhhereer in the control unit 16 , as well as in dependence on a speed of the internal combustion engine 10 educated.

Subsequently, in the step 26 the formation of a corrected cylinder-individual value Qcorr wherein the corrections are made by a known per se idle control and / or truancy. The idle control regulates a certain idle speed when the accelerator pedal is not actuated. The rider control ensures uniform injection quantities on all cylinders, thus improving smoothness and emissions. In step 28 becomes the corrected value Qcorr possibly to k partial injection quantities qk split by Qcorr is shared by k.

The partial injection quantity qk represents a reference quantity determined for the multiple injection pattern with k partial injections. As will be explained in more detail below, the actually injected fuel quantity may deviate from this reference quantity.

In step 30 will be the quantities qk a driving time AD associated with the injection actuator 20 is to drive to the desired quantities qk to dose. The assignment is done with the help of a in the control unit 16 deposited context, which in the 2 through the links next to the step 30 illustrated characteristic Q_EDC is represented. In the illustrated embodiment, the relationship is a straight line, the increasing drive times AD larger fuel quantities qk assigns and vice versa, wherein for dosing a fuel quantity Qk> 0 a Mindestansteuerdauer AD_0 is required.

As previously described, will be a lot qk predetermined and suitable for a control period AD as the value of the inverse function Q_EDC ^-1 for the crowd qk educated. Subsequently, in the step 32 at the correct angular position, the output of the driving time AD to the injection actuator 20 before the program enters the main program 22 branches back. Depending on the values k, single injections (k = 1) or multiple injections (k> 1) result.

3 shows the characteristic Q_EDC as they are in new condition of the injection system 12 in the control unit 16 stored together with a characteristic Q_ist how it actually results as a result of manufacturing tolerances and / or how they are due to aging of the injection system 12 including the injection actuator 20 over time. QLL is an amount of fuel, which by means of idle control and / or restraint control in a fixed idle operating point, ie at a specific torque request and a certain idle speed, as a result of the actual characteristic Q_ist actually injected. The control unit gives this 16 for a single injection 31 the drive time AD 1 out.

Due to the control unit 16 stored characteristic Q_EDC goes the control unit 16 however, assume that with the driving time AD1 the smaller injection quantity QL1 is dosed. There QL1 around the mistake F is smaller than QLL , are all calculations of the controller 16 that on the crowd QL1 based, also with this error F afflicted. This mistake F represents the deviation already mentioned and is also used in the following as a synonym for the deviation. While the total injection amount for a combustion process at idle is still corrected by the idling control, z. B. small pilot injections in other operating points of the internal combustion engine 10 Under certain circumstances, no separate regulatory intervention. For these small injection quantities, it is therefore of particular importance that the control unit 16 the actual matching between injected fuel quantity Q and driving time AD used.

The effect of a mismatch becomes clear even if one makes a transition from the already considered single injection of the set QLL considered on a multiple injection with which the internal combustion engine 10 to be operated at the same operating point. Without knowledge of the error F becomes the controller 16 output an injection pattern with n sub-injections, each of which has a drive duration Adn *. Adn * results as Adn * (QL1 / n) from the inverse function Q_EDC ^-1 the characteristic stored in the control unit Q_EDC , Because of the actual valid characteristic Q_ist However, with the drive time Adn *, not the calculated amount QL1 / n but the larger amount F + QL1 / n.

In the presentation of the 3 is n = 5. In the illustrated parallel position of Q_ist and Q_EDC Thus, each partial injection has the error F so that the sum of the five split injections has a quantity error that is five times greater than the error F the single injection. The control unit 16 must compensate for the fivefold greater error by reducing the Ansteuerverdauern the partial injections by a control process, which requires a disturbing transient. The injection pattern 33 shows the result of the control process, after the to be kept constant for a given operating point total injection quantity QLL per combustion chamber filling was divided into n = 5 partial injections, each with an injection quantity QLL / n actually injected.

To the mistake F to determine quantitatively, sees the above-mentioned DE 103 43 759 A1 The following: First, the amount required for the single injection QL1 saved. After switching to the multiple injection with n partial injections partial injection quantities are given the assumption of a constant operating point QLL / n set to which Ansteuerverdauern AD n belong. The control unit 16 knows in addition to the driving time AD1 and quantity QL1 the driving time adn and quantity QLn / n of a single one of n split injections and the number n of split injections. The quantity error F then turns to $$\text{F}=\left(\text{QL1}-\text{QLn}\right)/\left(\text{n}-1\right)$$

Where QLn / n is not the nth part of QL1 but the fictitious amount of a partial injection of the multiple injection 33 , which in the control to the constant operating point of the characteristic curve Q_EDC results. This calculation of the error F based on the premise that the quantity error F Additive nature is, so the characteristics Q_ist and Q_EDC run parallel. The parallelism must be particular over the entire distance DAD extend on the drive duration axis between the drive time AD n a partial injection and the driving time AD1 the large single injection is located.

The inventors have recognized that such parallelism can not be presupposed per se when large portions of the drive duration axis are used to determine the deviation. Any deviation from parallelism has an inaccuracy in determining the error F result.

By contrast, the invention uses the place of a large single injection 31 and a multiple injection 33 a first multiple injection pattern 35 and a second multiple injection pattern 37 to determine the error F ,

4 shows the conditions with the same characteristics Q_ist and Q_EDC like in the 3 in the event that the injection quantity QLL is injected with a first multiple injection pattern with x = 5 split injections and with a second multiple injection pattern with y = 4 split injections. QLL / x is the amount injected with a single split injection of the first multiple injection pattern at the constant operating point.

F is the additive error and QLx / x is the quantity supposed to be in the controller 16 stored characteristic Q_EDC is injected with a single partial injection of the first multiple injection pattern.

Then that is x times the sum of QLx / x and the error F equal to QLL, so $$\text{QLX}+\text{x}*\text{F}=\text{QLL}$$

Similarly, QLL / y is the amount injected with a single split injection of the second multiple injection pattern at the constant operating point. F is again the additive error and QLy / y is the quantity supposed to be in the controller 16 stored characteristic Q_EDC is injected with a single partial injection of the second multiple injection pattern.

Then y-fold is the sum of QLy / y and the error F equal to QLL, so $$\text{QLY}+\text{y}*\text{F}=\text{QLL}$$

In sum, with both multiple injection modes, the same amount QLL injected per combustion chamber filling. In this case, the equality results when the operating point of the internal combustion engine 10 is kept constant by a control process independent of the injection pattern used.

If the index i is a specific cylinder of the internal combustion engine 10 numbered, applies to the first multiple injection according to: $$\text{QLX}\text{i}=\text{QLL}\text{i}-\text{x * F}$$

The same applies for the second multiple injection pattern: $$\text{QLY}\text{i}=\text{QLL}\text{i}-\text{y}*\text{Fi}$$

Subtracting equations (4) and (5) and shifting the result to the cylinder-individual error Fi then returns the following quantitative result for the error Fi : $$\text{Fi}=\left(\text{QLX}\text{i}-\text{QLY}\text{i}\right)/\left(\text{y}-\text{x}\right)$$

The numbers x and y of the split injections of the first multiple injection pattern and the second multiple injection pattern differ at least by the value 1 ,

The cylinder-individual EDC quantities QLX . i and QLY, i are determined with the aid of the idling and running inhibitor. In this case, the control periods set for setting a stable idling operating point with the stored data of the context stored in the control unit are (fictitious) Q_EDC Expressed in terms of quantities. When shifted by a drift characteristic differ Q_EDC Amounts of the actually injected quantities by the constant assumed proportion Fi.

With the well-known characteristic drift, or the error Fi in the small amount range, the desired pre-injection amount can be adjusted by appropriate adjustment of the control period.

The width of the area DAD over which the characteristics Q_EDC and Q_ist for a reliable error detection to be parallel, when using multiple multiple injection pattern is much smaller than when using a large single injection and a multiple injection pattern. This reduces the risk that the underlying premise of parallelism will be violated and thus corrected uncertainly. In other words, the use of two multiple injection modes results in a smaller range on the drive duration axis over which the characteristics curve Q_ist and Q_EDC have to run parallel.

The probability that the characteristic curves are actually parallel there is greater than when a larger range is considered on the actuation duration axis. Therefore, the reliability of the determination of the error increases F when using two multiple injection patterns instead of a single injection pattern and a multiple injection pattern. A further advantage is that in the use of the first multiple injection pattern and the second multiple injection pattern, the quantity difference between the new state and the aged state, or between the desired state and the actual state is multiplied. In both cases, this increases the deviation between the reference variable and controlled variable of the control intervention of the idling control and / or restraint control, whereby the sensitivity with respect to a crowd drift is basically increased.

5 shows an embodiment of a method according to the invention. The step represents 22 that already in connection with the 2 mentioned main program HP for controlling the internal combustion engine 10 , Is a constant operating point BP = constant before, becomes in one step 34 a number k = x predetermined by partial injections of a first multiple injection pattern. Subsequently, the steps become 24 to 30 out 2 done with this value for k, so in the step 36 a driving time ADx output for k = x and in one step 38 together with the reference quantity QLX is stored. With this drive time ADx will the crowd QLL / x actually metered, the controller due to the characteristic Q_EDC However, that only assumes a lot QLX / x is dosed.

Then follow the steps 34 to 38 appropriate steps 40 to 44 for a number k = y performed by partial injections of a second multiple injection pattern. In step 46 then becomes the mistake F determined according to the equation (6) before the program about brands A . B into the main program of the step 22 Branched back. For determining individual cylinder errors Fi the procedure is carried out separately for each cylinder.

In a further refinement, the previously determined errors are compensated F and or Fi , This is done from the specific errors F and or Fi and in the control unit 16 stored characteristic Q_EDC_alt a new characteristic Q_EDC newly formed. The formation takes place by adding F or Fi to Q_EDC as it is by the step 48 is clarified. For compensation, the method is carried out at larger, predetermined intervals. In a preferred embodiment, the predetermined distance results from a completed between two bushings route of z km, where z is of the order of 1000.

Claims (9)

Method for determining the deviation (F) of the actual injection quantity from a predetermined reference quantity of a fuel injection system (12) of an internal combustion engine (10) having at least one controllable injection actuator (20) in which the injection quantity required at a specific operating point of the internal combustion engine (10) per combustion event (QLL) is injected for various combustion processes with different injection patterns, and wherein the deviation (F) is determined from a comparison of the reference quantity (QLx / x) formed in a first injection pattern with the reference quantity (QLy / y) formed in a second injection pattern characterized in that the first injection pattern (35) comprises x single injections and that the second injection pattern (37) comprises y single injections, where x and y are each greater than one, and x is not equal to y. Method according to Claim 1 Characterized in that the comparison is directly given to the control unit (16) reference quantities (QLX / x, QLY / y) based or indirectly to the reference amounts (QLX / x, QLY / y) corresponding actuation periods (ADx, ADy) is based. Method according to Claim 1 or 2 , characterized in that the determined deviation (F) in the further operation of the internal combustion engine (10) in the formation of reference quantities and / or Ansteuerdauern is compensated. Method according to Claim 3 , characterized in that it is repeated at predetermined intervals. Method according to Claim 1 or 2 , characterized in that the difference of x and y is equal to 1. Method according to Claim 1 or 2 , characterized in that the specific operating point is defined by constant values of the speed and the torque. Method according to one of the preceding claims, characterized in that the individual injection quantities (QLL / x, QLL / y) of an injection pattern (35, 37) are the same size. Control unit (16) for determining the deviation (F) of the actual injection quantity from a predetermined in the control unit (16) reference quantity of a fuel injection system (12) of an internal combustion engine (10) with at least one controllable injection actuator (20), wherein the control device (16) is established to form and output different injection patterns for the injection quantity (QLL) required for each combustion event at a particular operating point of the internal combustion engine (10) for each combustion event, and wherein the controller (16) comprises means for comparing the deviation (F) from a comparison of determining a reference quantity (QLx / x) formed with a second injection pattern, characterized in that the control unit (16) is designed such that it contains the first injection pattern (35) with x individual injections and forming second injection patterns (37) with y single injections and excl ibt, where x and y are each greater than one and where x is not equal to y. Control unit (16) to Claim 8 , characterized in that it is adapted to a method according to one of Claims 2 to 7 perform.

Images