DE102006007365B3 - Method for controlling and regulating an internal combustion engine, involves setting of minimum pressurization level from maximum individual accumulator pressure in first step - Google Patents

Abstract

Method involves setting of minimum pressurization level (pE(1)) from the maximum individual accumulator pressure (pEMax) in the first step and testing (pE(1) is less than pE(MAX)), whether first pressure value (p1) as well as second pressure value (p2) were included. In further step a further pressure level is determined. The method corresponding to the further step is set as long till (n+1) pressure level (pE (n+1)) is recognized in single or without pressure value and the injection is calculated on dependence of the first pressure value as well as the second pressure value of the n pressure-value (pE).

Description

The The invention relates to a method for controlling and regulating a Internal combustion engine according to the preamble of claim 1.

at an internal combustion engine determine the start of injection and the end of injection decisively the goodness the combustion and the composition of the exhaust gas. To the legal To comply with limit values, these two parameters are usually controlled by an electronic control unit. Because between the Energization start of the injector, the needle stroke of the injector and the actual Injection start a time lag occurs, is different the actual injection start from the target injection start. This causes unequal exhaust gas values for and the same operating point. For the At the end of injection, this applies accordingly.

In practice, to detect the actual start of injection, the armature impact at the injector is detected via a change in the PWM signal. Such a solution is for example from the DE 42 37 706 A1 known. Another way to detect the actual start of injection is to measure the pressure profile in the rail in a common rail system. By means of the differentiation of the pressure curve, the significant changes, ie the extreme values, are interpreted as the beginning and the end of the injection. The filtering of the measured pressure curve is absolutely necessary since, for example, the injection frequency and the delivery frequency of the high-pressure pump are present as disturbance variables on the pressure signal. However, the filtering of a signal causes a significant temporal offset from the raw signal. Such a solution is for example from the DE 31 18 425 A1 known.

Also the DE 197 40 608 A1 deals with the problem as in a conventional common rail system from the measured, noise-affected rail pressure, the fuel injection-related characteristics, such as the Einspitzmenge and the injection duration, can be determined. The method consists in that, in a first step, the fuel pressure profile in the rail is measured and stored. In a second step, this raw signal is then sampled with a specific resolution, low-pass filtered, a predetermined time window selected and the pressure values normalized. In a third step, this pressure profile pattern is supplied as an input vector to a neural network, via which ultimately the fuel injection-related parameters are determined.

From the DE 103 44 181 A1 A method is known for determining a virtual start of injection from an injection end in a common rail system with individual memories for fuel storage. The virtual injection start is calculated from the measured injection end via a mathematical function. Since the raw values of the individual accumulator pressure are used in this method, the interfering frequencies present in the system are superimposed on the measurement signal. This can cause a misinterpretation of the end of injection and cause a virtual start of injection deviating from the ideal state.

From the DE 10 2004 006 896 A1 is a test method for evaluating an injector in a common rail system with individual memories known. In this method, an injection end deviation from a desired-actual comparison of the injection ends and an injection start deviation from the nominal-actual comparison of the injection starts are calculated. The injector is rated as error-free if the start of injection and the end of injection are within the respective tolerance band. Otherwise, a diagnosis entry takes place and the control parameters for controlling the injector, for example the start of energization, are adjusted accordingly. Since the start of injection according to the in the DE 103 44 181 A1 calculated method is calculated from the injection end, the previously described problem of misinterpretation also occurs here.

The invention is based on the object, from the DE 103 44 181 A1 to develop known methods for a common rail system with individual memories with improved detection of the end of injection.

The The object is solved by the features of claim 1. The Embodiments are shown in the subclaims.

The Invention proposes a recursion method for determining the end of injection, which the stored raw values of the individual memory pressure during a Measuring interval based. The invention is based on the knowledge that that in the measuring interval the pressure curve in the individual memory falls a falling Branch, injection activated, and a rising branch, injection disabled. While an injection thus exist for each pressure level a first Pressure reading on the falling branch and at least a second pressure reading on the way up. Will only be a single or none Pressure measured detected, so the end of injection must be present or has taken place.

The method consists of a first step and a further step, which is repeated until the detection of the end of injection. In the first step, starting from a maximum Single memory pressure a first, lower pressure level fixed and checked whether a first pressure reading and a second pressure reading to the first pressure level were detected. At a maximum individual storage pressure of 1800 bar and a step size of 10 bar, the first pressure level is 1790 bar. In the next step, a further pressure level is set, which is reduced compared to the previous one. Thereafter, it is also checked whether a first and a second pressure measured value have been detected for the further pressure level. If the further step corresponds, for example, to the second step, then the further pressure level, ie the second pressure level, is set to 1780 bar.

The The process continues according to the next step until at a (n + 1) -th pressure level only a single or no pressure reading is detected anymore. The injection end then becomes dependent on the first pressure reading and the second pressure reading of the nth pressure level, So the previous pressure level, calculated. For example, by the average of the first pressure reading and the second pressure reading is calculated.

Around making the process less susceptible to disturbance is in one Design provided that the first and the other pressure levels correspond to a tolerance band. For several within the tolerance band detected first pressure readings is a representative first pressure reading certainly. For several within the tolerance band detected second Pressure readings are selected analogously. Opposite one A procedure in which all measured values are filtered offers these Approach still a time gain.

In an embodiment is for more accurate determination of the end of injection provided that in the range between the n-th pressure level and the (n + 1) -th pressure level, the procedure according to the other Step with a smaller increment, for example 1 bar, repeated becomes.

In One embodiment is the calculation of a representative virtual injection start shown. This is calculated by mean value formation from three virtual injection starts. The first virtual start of injection calculated from the end of injection as well as in the measuring interval determined first pressure readings on a regression line. The second virtual start of injection is calculated from the end of injection and the first pressure reading of the first pressure level via a Just. The third virtual start of injection is calculated the injection end and the first pressure measured values determined in the measuring interval via a Parabola. By this measure becomes the influence of disturbances on the determination of the virtual start of injection is further reduced. The further control and regulation of the internal combustion engine takes place based on the representative virtual injection start and the previously calculated injection end.

Next a timely determination of the start of injection and end of injection offers the invention as a further advantage a precise, injector close Illustration of the hydraulic properties of the injection system during the Injection. Under hydraulic properties of the system are the characteristics of the Injektors, the leakage, the needle stroke and the behavior of the individual memory to understand.

In The drawings show a preferred embodiment.

It demonstrate:

1 a system diagram;

2 an injection;

3 a diagram for determining the end of injection;

4 a diagram for determining the representative virtual injection start;

5A . 5B a program schedule.

The 1 shows a system diagram of an electronically controlled internal combustion engine 1 , In this case, the fuel is injected via a common rail system. This includes the following components: a low-pressure pump 2 for fuel delivery from a fuel tank 3 , a suction throttle 4 for determining a volume flow, a high-pressure pump 5 to promote the fuel under pressure increase in a rail 6 , Single memory 7 for buffering the fuel and injectors 8th for injecting the fuel into the combustion chambers of the internal combustion engine 1 ,

The common rail system with individual memories differs from a conventional common rail system in that the fuel to be injected from the individual memory 7 is removed. The supply line from the rail 6 to the individual memory 7 is designed in practice so that a feedback of noise in the rail 6 is dampened. Just as much fuel flows out of the rail during the injection break 6 after that the single memory 7 is refilled at the beginning of the injection. The hydraulic resistance of the single memory 7 and the supply line are matched, ie the connecting line from the rail 6 to the individual memory 7 has the highest possible hydrauli resistance. In a conventional common rail system without single memory, the hydraulic resistance between the rail 6 and the injector 8th be as low as possible to achieve an unimpeded injection.

The operation of the internal combustion engine 1 is controlled by an electronic control unit (ADEC) 9 regulated. The electronic control unit 9 includes the usual components of a microcomputer system, such as a microprocessor, I / O devices, buffers and memory devices (EEPROM, RAM). In the memory modules are those for the operation of the internal combustion engine 1 Relevant operating data in maps / curves applied. This is calculated by the electronic control unit 9 from the input variables the output variables. In 1 The following input variables are shown by way of example: a rail pressure pCR, which is determined by means of a rail pressure sensor 10 is measured, a speed signal nMOT the internal combustion engine 1 , Pressure signals pE of the individual memory 7 and an input ON. For example, the input quantity EIN subsumes the charge air pressure of a turbocharger and the temperatures of the coolant / lubricant and of the fuel.

In 1 are the output variables of the electronic control unit 9 a signal PWM for controlling the suction throttle 4 and an output OUT shown. This is representative of the other control signals for controlling and regulating the internal combustion engine 1 , For example, an injection quantity ve for representing a desired torque in a torque-based control.

In 2 an injection is shown. The ordinate shows the individual accumulator pressure pE in bar and the abscissa the crankshaft angle Phi in degrees. Within the diagram is a measured pressure curve pE = f (Phi), a measuring interval MESS in the angular range of 300-500 degrees crankshaft angle Phi, and several isobars drawn as dashed lines. The maximum static single-store pressure pE (MAX) for this system is 1800 bar.

The Injection end is over determines a recursion procedure which is based on the stored raw values of the individual memory pressure pE is based on the measuring interval MESS. The invention is the realization based on that in the measuring interval MESS the pressure curve in the individual memory a falling branch, injection activated, and a rising one Ast, injection disabled, has. So that exists for everyone Pressure level is a first pressure reading on the falling branch and at least a second pressure reading on the ascending branch. Will only one more single or no pressure reading detected, so must the injection end be present or have taken place.

The method for determining the tipping end is as follows:
In a first step, starting from the maximum static individual accumulator pressure pE (MAX), a first, lower pressure level pE (1) is determined. With a step size dp1 of 10 bar, the first pressure level is calculated to pE (1) = 1790 bar. Thereafter, it is checked whether a first pressure reading p1 and a second pressure reading p2 have been detected at the first pressure level pE (1). This is in the 2 the case. In a further step, a further pressure level is then determined and it is checked whether a first p1 and second pressure value p2 have been detected for this further pressure level. If the further pressure level is, for example, the second pressure level pE (2), then this is calculated from the first pressure level pE (1) = 1790 bar minus the step size dp1 = 10 bar to pE (2) = 1780 bar. As an example, an eighth pressure level pE (8) = 1720 bar is plotted. At this pressure level pE (8), a first pressure measurement p1 and a second pressure measurement p2 were also detected.

The process is continued until only one or no pressure reading is detected. In the 2 For example, an (n + 1) -th pressure level pE (n + 1) is plotted with only one pressure reading p1. In this case, from the previous pressure level pE (n), the first pressure reading p1 and second pressure reading p2 are used for determining the injection end of the SE. The injection end SE is calculated from the average of the first p1 and second pressure reading p2 of the nth pressure level pE (n). If, for example, at the nth pressure level the first pressure measurement p1 is defined by the value pair pE = 1650 bar and crankshaft angle Phi = 370 degrees and the second pressure measurement p2 is defined by the value pair pE = 1650 bar and Phi = 380 degrees, then the end of injection SE is calculated pE = 1650 bar and Phi = 375 degrees.

The procedure can be like in 3 clarify. Of the 3 was based on the example described above, ie at (n + 1) -th pressure level pE (n + 1), only a single pressure reading p1 was detected, ie the end of injection must have taken place. Starting from the previous pressure level, here pE (n), the recursion procedure is repeated according to the method described in 2 described further step with a smaller step size dp2 repeated. For the representation in 3 a step size dp2 = 1 bar was used.

The procedure is as follows:
At the nth pressure level pE (n), for example pE (n) = 1660 bar, a first pressure reading p1 and a second pressure reading p2 detected. Thereafter, in the next step, the pressure level is reduced by the step size dp2 to 1649 bar and checked again whether a first p1 and second pressure value p2 were detected. This is in 3 the case. The pressure level is reduced in 1-bar steps until at a pressure level of 1646 bar only the pressure reading p1 is detected, here the point A. The injection end SE is then calculated from the previous pressure level 1647 bar, from the mean value of first pressure reading p1 (1647 bar / 376.2 degrees) and second pressure reading p2 (1647 bar / 377 degrees) with the value pair injection end SE (1647 bar / 376.6 degrees).

The 4 shows a diagram for determining the representative virtual injection start SBvr. Shown is the falling branch of the individual memory pressure curve pE = f (Phi) during an injection. This graph is based on the same raw measurements as in 2 and in 3 , The representative virtual injection start SBvr is calculated from a first, second and third virtual start of injection.

Of the first virtual injection start SBv1 is based on the end of injection SE and the first pressure measured values determined in the measuring interval MESS p1 over a regression straight line RGE (solid line) is calculated. Of the first virtual injection start SBv1 then results from the intersection the isobars to the maximum single-store pressure pE (MAX) and the Regression line RGE. The second virtual start of injection SBv2 is determined by the injection of the SE and the first pressure reading p1 of the first pressure level pE (1) = 1790 bar over a straight line GE (dot-dashed line) Line). The second virtual start of injection SBv2 results then from the intersection of the isobars to the maximum single-reservoir pressure pE (MAX) and the line GE. The third virtual injection start SBv3 is determined by the injection end SE and the measuring interval MESS determined first pressure readings p1 via a parabola PAR (dash-dotted two Line). The third virtual start of injection SBv3 results then from the intersection of the isobars to the maximum single-store pressure pE (MAX) and the parable PAR. The representative virtual Start of injection SBvr is then over Averaging the three virtual injection starts SBv1, SBv2 and SBv3 calculated.

The Definition of the first, second and third virtual start of injection can also over other mathematical functions, e.g. a cubic function. Of course The number of virtual injection starts can also be changed.

In 5 3, a flowchart for the subroutine for calculating the injection end SE and the representative virtual injection start SBvr is shown. The 5 consists of the subfigures 5A and 5B , Reference numerals S1 to S5 correspond to the first step. The reference symbols S6 to S12 correspond to the further step.

At S1 in 5A the individual accumulator pressure values pE are detected and stored during the measuring interval MESS, ie each raw measured value is stored as a pair of values of a pressure value pE and of a crankshaft angle Phi. At S2, starting from the maximum individual accumulator pressure pE (MAX), a first, lower pressure level pE (1) is established by subtraction with the increment dp1. Thereafter, it is checked at S3 whether a first pressure reading p1 and a second pressure reading p2 to the first pressure level pE (1) have been detected. If no or only one pressure measurement value can be determined, query S3: no, then there is an error, ie a diagnosis entry is made and an alternative function is activated, S4.

Become a first p1 and second pressure reading p2 detected, query S3: yes, in S5 the run variable n is set to two. After that a further pressure level is calculated, which is opposite to the previously reduced by the increment dp1, S6. At S7 becomes checked, whether a first p1 and a second pressure reading p2 to the further pressure level were recorded. If this is the case, query S7: yes, then at S8 the run variable n increased by one and branched to the program point A.

If no or only one pressure measurement value is detected at S7, query S7: no, then this pressure level is stored at S9. In 5A this is exemplified as (n + 1) pressure level pE (n + 1). At S10, the previous pressure level pE (n) is decreased by a step size dp2, for example 1 bar, and at S11 it is checked whether there is a first p1 and second pressure value p2 at this pressure level. If this is the case, query S11: yes, it is branched to the program point B and the program continues with S10, ie the pressure level is again reduced by the step size dp2. If it is determined at S11 that only one or no pressure measurement value has been found, query S11: no, the injection end SE is calculated at S12 from the pressure level last detected with two pressure measurement values. In this case, the mean value MW is calculated from the first p1 and second pressure measurement p2. The injection end SE then corresponds to this mean value MW. At S13, a branch is made to the subroutine calculation of the representative virtual injection start SBvr. This is in the 5B shown.

There is a first virtual injection at S1 SBv1 is calculated on the basis of the injection end SE and the first pressure measurement values p1 determined in the measurement interval MESS via the regression line RGE. Thereafter, a second virtual injection start SBv2 is calculated via the injection end SE and the first pressure measurement value p1 of the first pressure level pE (1), for example 1790 bar, via the straight line GE, S2. At S3, a third virtual injection start SBv3 is calculated based on the injection end SE and the first pressure measurement values p1 determined in the measurement interval MESS via the parabola PAR. The representative virtual injection start SBvr is then determined from the average value of the three virtual injection starts SBv1, SBv2 and SBv3, S4. Thereafter, the subroutine is finished and it becomes S13 in the 5A and returned to the main program, in which the representative virtual injection start SBvr and the injection end SE are used as relevant for the further control of the internal combustion engine.

at the description of the method was based on raw values, which are related to the crankshaft angle Phi. Hence the values the injection end SE and the representative virtual injection start SBvr also related to the crankshaft angle Phi. As alternative can this Characteristics of time be. In this case, the reference to the crankshaft angle Phi in the figures as a reference to the time to understand.

In The figure description, the individual pressure levels were through a fixed numeric value, for example the first Pressure level pE (1) by the numerical value 1790 bar. Instead of a Fixed value, the pressure level can be characterized by a tolerance band be. The specified numerical value then corresponds to the middle of the range of the tolerance band, for example pE (1) = 1790 bar ± 5 bar. The reference to a corresponding pressure level in the figure description is therefore also to be understood as a reference to a tolerance band. Becomes within the tolerance band several first pressure readings p1 is determined, then from these a representative first pressure reading for this Tolerance band determined. For the second pressure readings apply the same approach.

• – das berechnete Einspritzende und der repräsentative virtuelle Einspritzbeginn sind robust gegenüber Störgrößen, wodurch der Regelkreis zur Regelung des Einspritzbeginns und Einspritzendes stabilisiert wird;
• – eine zylinderindividuelle Diagnosewert-Aussage und Regelung ist möglich;
• – die hydraulischen Eigenschaften des Einspritzsystems im Bereich der Einspritzung sind präziser;
• – der messtechnische Aufwand während der Entwicklungsphase der Brennkraftmaschine wird verringert.

From the description, the following advantages result for the method according to the invention:

• The calculated injection end and the representative virtual start of injection are robust against disturbances, whereby the injection start and end control loop is stabilized;
• - A cylinder-specific diagnosis value statement and control is possible;
• - The hydraulic properties of the injection system in the injection area are more precise;
• - The metrological effort during the development phase of the internal combustion engine is reduced.

1

Internal combustion engine

2

Low pressure pump

3

Fuel tank

4

interphase

5

High pressure pump

6

Rail

7

Single memory

8th

injector

9

electronic control unit (ADEC)

10

Rail pressure sensor

Claims (8)

Method for controlling and regulating an internal combustion engine ( 1 ) with a common rail system together with individual memories ( 7 ) for temporary storage of fuel, in which a single storage pressure (pE) during a measurement interval (MESS) is detected and stored and a significant change in the individual storage pressure (pE) is interpreted as the injection end (SE), characterized in that in one first step, starting from a maximum individual accumulator pressure (pE (MAX)), a first, lower pressure level (pE (1)) is set (pE (1) <pE (MAX)), it is checked whether at least a first pressure reading (P1) and at least a second pressure measurement (p2) to the first pressure level (pE (1)) were detected, in another step, a further pressure level is set, which is reduced compared to the previous pressure level, it is checked whether at least a first (p1 ) and at least a second pressure reading (p2) were detected to the further pressure level, and the method is continued according to the further step until at an (n + 1) -th pressure level (pE (n + 1)) only e in individual or no pressure measurement (p1, p2) is detected and the injection end (SE) in response to the first pressure reading (p1) and the second pressure reading (p2) of the n-th pressure level (pE (n)) is calculated. Method according to claim 1, characterized in that that the first pressure level (pE (1)) and the other pressure levels correspond to a tolerance band, with several detected within the tolerance band first pressure readings (p1) represent a representative first pressure reading is determined and at several within the tolerance band detected second Pressure readings (p2) a representative second Pressure reading is determined. A method according to claim 1 or 2, characterized in that in the area between the n-th pressure level (pE (n)) and the (n + 1) -th pressure level (pE (n + 1)), the process is repeated according to the next step with a smaller step size (dp2). Method according to claim 3, characterized that the end of injection (SE) is calculated by being found last Pressure level with two pressure readings from the first pressure reading (p1) and the second pressure reading (p2) of this pressure level Mean value (MW (p1, p2)) is determined. Method according to claim 4, characterized in that a first virtual start of injection (SBv1) based on the end of injection (SE) and the first pressure readings determined during the measurement interval (MESS) (p1) over a regression line (RGE) is calculated, a second virtual Start of injection (SBv2) based on the end of injection (SE) and the first pressure reading (p1) of the first pressure level (pE (1)) via a Just (GE) is calculated, a third virtual injection start (SBv3) based on the end of injection (SE) and in the measuring interval (MESS) calculated first pressure readings (p1) via a parabola (PAR) and from the three virtual injection starts (SBv1, SBv2, SBv3) a representative virtual injection start (SBvr) is determined. Method according to claim 5, characterized in that that the first (SBv1), the second (SBv2) and the third virtual Start of injection (SBv3) from the intersection of the isobars to the maximum Single storage pressure (pE (MAX)) and the respective mathematical Function (RGE, GE, PAR) is calculated. Method according to Claim 6, characterized that the representative virtual injection start (SBvr) over mean value formation of three virtual injection starts (SBv1, SBv2, SBv3) is calculated. A method according to claim 7, characterized in that the representative virtual injection start (SBvr) and the injection end (SE) for the further control and regulation of the internal combustion engine ( 1 ) is set as determining.