DE102008059609A1 - Method for determining angle position of combustion of internal combustion engine, involves determining parameter of model function, and determining characteristics of combustion of engine depending on parameter - Google Patents

Abstract

The method involves detecting a sequence of rotational speed values during operation of an internal combustion engine, where the sequence of the rotational speed values is approximated as a periodic function with a model function i.e. Fourier series function. A parameter i.e. Fourier coefficient, of the model function is determined, and characteristics i.e. angle position, of combustion of the engine are determined depending on the parameter. A scalar product is computed with one of the rotational speed values of the sequence of rotational speed values.

Description

The The present invention relates to a method for determining Characteristics of the combustion of an internal combustion engine according to the features of claim 1.

From the DE 10 2006 026 380 A1 it is already known to determine parameters of combustion of an internal combustion engine. For this purpose, a sequence of rotational speed values is detected, transformed into the frequency range, and a current ignition frequency component in the frequency range is determined. The current ignition frequency component is further compared with an ignition frequency component determined in advance using a reference internal combustion engine. On the basis of a difference between the actual ignition frequency component and the ignition frequency component determined in advance and by means of an estimation characteristic, a parameter of the combustion of the internal combustion engine is determined, for example the position of the combustion in a cylinder.

According to this Stand, however, it is a disadvantage that only the ignition frequency component as an information source for the determination of parameters the combustion is used. In particular, it can be assumed that only in the range of lower speeds sufficiently accurate parameters The combustion can be determined because in the area higher speeds due to the moments of inertia the oscillating masses and dynamic processes at the Crankshaft, such as torsion, also higher frequency components play a role.

task

It It is an object of the present invention to provide a method for determining Characteristics of the combustion of an internal combustion engine to provide that does not have these disadvantages, so that also in the range of higher speeds precise characteristics combustion can be determined.

solution

These The object is achieved according to the invention that during operation of an internal combustion engine a sequence of speed values is detected, the sequence of Speed values as a periodic function with a model function is approximated, whereby parameters of this model function are determined and depending on the parameter characteristics the combustion are determined. The sequence of speed values is preferably over the segment of at least one cylinder but preferably over the segments of all cylinders of the internal combustion engine so that is preferred for at least one for all cylinders of the internal combustion engine characteristics combustion can be determined. According to the invention the model function a Fourier series, whose parameters are the Fourier coefficients are. The Fourier coefficients are inventively during the operation of the internal combustion engine by the calculation a scalar product over selected rows a matrix with the individual speed values from the detected Calculated sequence of speed values. This matrix is inventively the inversion of a system of equations in the context of development the internal combustion engine determined. This equation system becomes according to the model function shown as Fourier series described in connection with the detected sequence of speed values. A characteristic of the combustion is, for example the time or angular position, at which 50% of the combustion chamber supplied energy, the situation at which the cylinder pressure is maximum or the indicated mean pressure. According to the invention advantageous is in this way also in the range of higher speeds an exact determination of combustion parameters possible because not only the ignition frequency component, but all frequency components as a source of information be used.

embodiment

Further advantageous embodiments of the present invention are the subsequent embodiment and the dependent To claim.

in this connection shows:

1 : the course of a sequence of speed values above the crank angle.

aufgefasst werden, wobei φ . für den einzelnen Drehzahlwert und der Index n für den jeweiligen Zylinder sowie der Wert in Klammern für den zugehörigen Kurbelwinkel in Grad Kurbelwinkel steht, also φ .₁(174) für den erfassten Drehzahlwert φ . in Umdrehungen pro Minute oder 1/s des ersten Zylinders n = 1 bei einem Kurbelwinkel von 174 Grad eines Segmentes von 180 Grad bei 4 Zylindern. Mit anderen Worten wird jedem Drehzahlwert φ . ein Drehwinkelwert φ zugeordnet.An internal combustion engine not shown has, as is well known, means for detecting the rotational speed of the crankshaft or camshaft. For example, by means of an inductive Sen Sors a sequence of teeth sampled, which are arranged on the flywheel of the internal combustion engine. For example, the gear comprises a series of 60 teeth, so that a tooth spacing of 6 degrees crank angle results. The internal combustion engine comprises at least one, but preferably a plurality of cylinders. For example, if the internal combustion engine includes 4 cylinders and the internal combustion engine operates according to the so-called four-stroke principle, every 180 degrees crank angle is a work cycle or combustion. This angular range of 180 degrees crank angle is also referred to as a cylinder segment. During the 180 degree crank angle of each one of the successive cylinder segments, a sequence of 30 rotational speed values can now be detected during operation of the internal combustion engine with a tooth spacing of 6 degrees crank angle. According to the invention, this sequence of rotational speed values of each cylinder segment of the exemplary four-cylinder internal combustion engine is continuously recorded during operation of the internal combustion engine. The sequence of 30 speed values can be used as a 30 × 1 column vector SV for each of the 4 cylinders

be understood, where φ. stands for the individual speed value and the index n for the respective cylinder as well as the value in brackets for the associated crank angle in degrees crank angle, ie φ. ₁ (174) for the detected speed value φ. in revolutions per minute or 1 / s of the first cylinder n = 1 at a crank angle of 174 degrees of a segment of 180 degrees at 4 cylinders. In other words, each rotational speed value φ. associated with a rotational angle value φ.

According to 1 is an example of the course of the sequence of speed values φ. shown above the crank angle, wherein in each case the detected rotational speed values φ. are shown for a cylinder segment of a particular cylinder in the form of boxes. According to 1 In an exemplary firing order of 1-3-4-2 of the internal combustion engine, the sequence of the cylinder segments of the cylinders 1, 3, 4, 2, 1 and 3 is shown in total, with the segment in each case being highlighted over 180 degrees of the cylinder 1 with boxes. The start or the end of the individual cylinder segments is according to 1 represented by individual points.

wobei φ . der jeweilige Drehzahlwert, A₀ / 2 der Gleichanteil, f_z die Zündfrequenz der Verbrennungskraftmaschine, φ der jeweils zu dem Drehzahlwert φ . zugehörige Drehwinkelwert, k die Motorordnung, n der Index für den jeweiligen Zylinder und A_k, B_k die Fourierkoeffizienten beziehungsweise die Parameter der Modellfunktion sind. Daraus ergibt sich, dass jedes Zylindersegment einen eigenen Satz Fourierkoeffizienten A_k, B_k beziehungsweise einen eigenen Satz Parameter der Modellfunktion hat, was aufgrund der unterschiedlichen zylindersegmentindividuellen Verläufe der Folge von Drehzahlwerten φ . auch sehr zweckmäßig ist. Mit anderen Worten wird ein Drehzahlmodell als Fourierreihe aufgestellt, das die Motorordnungen k als Frequenzen mit Grad Kurbelwinkel als unabhängige Variable, also als Frequenz in 1/Grad Kurbelwinkel, enthält. Dabei wird bevorzugt als erste Motorordnung k = 1 nicht das Arbeitsspiel 1/720 Grad Kurbelwinkel, sondern die Zündfrequenz f_z angesetzt, die bei einer Verbrennungskraftmaschine mit 4 Zylindern 1/180 Grad Kurbel winkel entspricht. Im Sinne der vorliegenden Erfindung entsprechen die weiteren Motorordnungen k > 1 dabei den Oberwellen der ersten Motorordnung k = 1, also Vielfachen der Zündfrequenz f_z, bei einer Verbrennungskraftmaschine mit 4 Zylindern ist dann die zweite Motorordnung k = 2 1/90 Grad Kurbelwinkel, die dritte Motorordnung k = 3 1/60 Grad Kurbelwinkel und so weiter. Erfindungsgemäß werden die Parameter der Modellfunktion, also die Fourierkoeffizienten A_k, B_k wie folgt bestimmt. Auf Grundlage der Folge von Drehzahlwerten φ . in Form des 30 × 1 Spaltenvektors SV und des Polynoms P kann folgendes Gleichungssystem GS in Matrixform aufgestellt werden:

hier exemplarisch für den ersten Zylinder n = 1 einer Verbrennungskraftmaschine mit 4 Zylindern bis zur achten Motorordnung k = 8. Das Gleichungssystem GS kann auch in der Form b = A·x (GS)dargestellt werden, wobei b dem 30 × 1 Spaltenvektor SV, die Matrix A der Verknüpfung des Sinus- und Kosinusanteils des Polynoms P mit der Zündfrequenz f_z, dem jeweils zu dem Drehzahlwert φ . zugehörigen Drehwinkelwert φ und der Motorordnung k sowie der Vektor x den Fourierkoeffizienten A_k, B_k beziehungsweise dem Gleichanteil A₀ / 2 entsprechen. Erfindungsgemäß können die Fourierkoeffizienten A_k, B_k nun durch Invertierung des Gleichungssystems GS folgendermaßen gebildet werden x = (ATA)–1AT·b (PI), also durch die Bildung einer Pseudoinversen PI des Gleichungssystems GS für bevorzugt jeden Zylinder. Aus der Pseudoinversen PI eines jeden Zylinders sind nun Zeilen auswählbar, die während des Betriebes der Verbrennungskraftmaschine zur Berechnung der Fourierkoeffizienten A_k, B_k erforderlich sind.According to the invention, the sequence of rotational speed values φ. according to the column vector SV as a periodic function approximated with a model function. For this purpose, according to the invention, a model function is used which corresponds to a Fourier series. Preferably, an approximation of the sequence of rotational speed values φ takes place. in the form of the trigonometric polynomial P for each cylinder of the internal combustion engine

where φ. the respective rotational speed value, A₀ / 2 the DC component, f _z the ignition frequency of the internal combustion engine, φ in each case to the rotational speed value φ. associated rotational angle value, k is the engine order, n is the index for the respective cylinder and A _k , B _{k are} the Fourier coefficients or the parameters of the model function. As a result, each cylinder segment has its own set of Fourier coefficients A _k , B _k or its own set of parameters of the model function, which due to the different cylinder segment-individual profiles of the sequence of rotational speed values φ. is also very appropriate. In other words, a speed model is set up as a Fourier series, which contains the engine orders k as frequencies with degrees crank angle as an independent variable, ie as a frequency in 1 / degree crank angle. In this case, preferably the first engine order k = 1 not the working cycle 1/720 degrees crank angle, but the ignition frequency f _{z is} used, which corresponds to 1/180 degrees crank angle in an internal combustion engine with 4 cylinders. For the purposes of the present invention, the other engine orders k> 1 correspond to the harmonics of the first engine order k = 1, ie multiples of the ignition frequency f _z , in an internal combustion engine with 4 cylinders then the second engine order k = 2 1/90 degrees crank angle, the third engine order k = 3 1/60 degrees crank angle and so on. According to the invention, the parameters of the model function, that is to say the Fourier coefficients A _k , B _k , are determined as follows. Based on the sequence of speed values φ. in the form of the 30 × 1 column vector SV and the polynomial P, the following equation system GS can be set up in matrix form:

Here, by way of example, for the first cylinder n = 1 of an internal combustion engine with 4 cylinders up to the eighth engine order k = 8. The system of equations GS can also be in the form b = A x (GS) where b is the 30 × 1 column vector SV, the matrix A is the combination of the sine and cosine components of the polynomial P with the ignition frequency f _z , in each case to the rotational speed value φ. associated rotation angle value φ and the motor order k and the vector x the Fourier coefficients A _k , B _k and the DC component A₀ / 2 correspond. According to the invention, the Fourier coefficients A _k , B _{k can} now be formed as follows by inverting the equation system GS x = (A T A) -1 A T · B (PI), that is, by the formation of a pseudoinverse PI of the equation system GS for preferably each cylinder. From the pseudoinverse PI of each cylinder lines are now selectable, which are required during operation of the internal combustion engine for calculating the Fourier coefficients A _k , B _k .

According to the invention, the Fourier coefficients A _k , B _k flow into a linear model for determining characteristic values of combustion KG according to equation (1), the values assigned to the Fourier coefficients A _k , B _{k being} generally denoted by F _n according to equation (1) are. KG = K 1 · F 1 + ... + K n · F n + K n + 1 (1)

If, for example, the Fourier coefficient A _{1 is to be included} in equation (1), if the value F _{1 is} assigned to it or if, for example, the Fourier coefficient B _{8 is to be included} in equation (1), the value F _{5 is} assigned to it. In other words, the assignment can be made arbitrarily.

The factors K _n are stored depending on the operating point for each cylinder in maps, for example, load and speed of the internal combustion engine and thus serve as operating point-dependent weighting of the individual Fourier coefficients A _k , B _k and F _n in the linear model to determine the characteristics of the combustion KG.

How many Fourier coefficients A _k , B _{k are} ultimately used in the method according to the invention for determining characteristic values KG of combustion of an internal combustion engine is freely selectable and should be considered taking into account the required storage requirements. It can be more or less. With a smaller number of Fourier coefficients, however, experience has shown that the approximation is rather poor, since over the entire operating range of an internal combustion engine, which is characterized by different speed and torque ranges, extremely fluctuating mechanical and thermal boundary conditions prevail. With a large number of Fourier coefficients A _k , B _k , the approximation or the model quality improves. However, 4 Fourier coefficients F ₁ ... F ₄ and 5 factors K ₁ ... K ₅ are preferably used for the linear model for determining characteristic values KG of the combustion of an internal combustion engine, in accordance with equation (2). KG = K 1 · F 1 + K 2 · F 2 + K 3 · F 3 + K 4 · F 4 + K 5 (2)

Which lines are used for calculating the Fourier coefficients A _k , B _k or F _n is inventively preferably determined in the context of the development of the respective internal combustion engine. For this purpose, for example, the characteristic variable of the combustion KG of an internal combustion engine, which characterizes the position of the combustion, ie the time or crank angle at which 50% of the energy supplied to the respective cylinder are converted, is determined and recorded on the basis of a cylinder pressure measurement for each individual cylinder of the internal combustion engine , In parallel, will be for the individual cylinder segments sequences of speed values φ. recorded and recorded. In other words, as part of a laboratory test on an internal combustion engine, sequences of rotational speed values φ. recorded and stored for the individual cylinder segments and recorded and stored in parallel for the individual cylinder segments associated working or combustion cycles sequences of pressure values in each cylinder, wherein an assignment of the sequences of rotational speed values φ. is possible from the sequences of the pressure values, for example by means of a combustion curve analysis of the crank angle at which 50% of the energy supplied to the respective cylinder are converted, so that each cylinder segment has a crank angle value, in which 50% of the energy supplied to the respective cylinder are converted, can be assigned. In the further course, the consequences of speed values φ. supplied to the equation system GS for the individual cylinder segments or determines the Fourier coefficients A _k , B _k or F _n by means of the pseudoinverse PI of the equation system GS. The Fourier coefficients A _k , B _k and F _n are then fed to the linear model according to equation (1) or equation (2), depending on how accurate the approximation should be, ie a linear model with more or less Fou rierkoeffizienten A. _k , B _k and F _n, respectively. Furthermore, equation (1) or equation (2) are supplied to factors K _n , which are determined by linear regression with the aid of the measurement data. Subsequently, the parameters of combustion KG are calculated on the basis of equation (1) or equation (2). In order to determine the Fourier coefficients A _k , B _k and F _n, respectively, which have the best correlation with the parameter of combustion KG, all combinations are calculated from the characteristic KG and calculated on the basis of equations (1) or (2) metrologically recorded characteristic of combustion KG formed and searched, for example, the combination with the least square error sum. In other words, the linear model according to equation (1) or (2) is set up with all possible combinations of all possible Fourier coefficients A _k , B _k and F _n and then looked at which combination has the smallest error compared with the measured reference data. Assuming 4 Fourier coefficients F ₁ ... F ₄ are used, then with all combinations of 4 out of all Fourier coefficients F _n the linear model is set up and looked according to equation (1) or (2), which combinations are the smallest error compared to have measured reference data, wherein the 4 Fourier coefficients F ₁ ... F _{4 are} then the most optimal for the respective cylinder. This procedure is performed individually for all cylinders. For the operation of an internal combustion engine, only the Fourier coefficients F ₁ ... F ₄ are then used which have the highest model quality, that is to say which have the best correlation between measured characteristic values of the combustion KG calculated using the linear model. Only these lines of the pseudoinverse PI, which have a determination of the Fourier coefficients F ₁ ... F ₄ with the highest model quality, are now used according to the invention for further processing, since the Fourier coefficients F ₁ ... F ₄ correspond to these lines. Accordingly, for the practical operation of an internal combustion engine, only these lines of the pseudoinverse PI, which have a determination of the Fourier coefficients F ₁ ... F ₄ with the highest model quality, are preferably stored in the control unit of the internal combustion engine for each cylinder and a selection takes place at runtime of these lines in the pseudoinverse PI. In particular, for the case where 4 Fourier coefficients F ₁ ... F _{4 are used} for the linear model, the 4 lines stored in the control unit per cylinder of the internal combustion engine are further processed in the pseudoinverse PI in such a way that during the transit time each Fourier coefficient F ₁ . ..F ₄ only a scalar product over a relevant line of the pseudoinverse PI with the speed values φ. the sequence of speed values φ. is calculated over a cylinder segment. In other words, a 4 × 30 matrix per cylinder corresponding to the pseudoinverse PI with the 30 × 1 column vector SV of the sequence of rotational speed values φ must be available at runtime. which then yields the current 4 Fourier coefficients F ₁ ... F ₄ for the respective cylinder segment as 4 × 1 column vector SV1. The calculation of a scalar product is carried out as generally known by component-wise multiplying the coordinates of the 30 × 1 column vector SV with the 4 rows of the pseudoinverse PI and then adding up the individual products.

Prefers the process according to the invention is combined with a well-known device for control and regulation run an internal combustion engine, at least a memory, a processor and appropriate electrical connections to sensors and actuators.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• - DE 102006026380 A1 [0002]

Claims (5)

Method for determining characteristics the combustion (KG) of an internal combustion engine, wherein during the operation of the internal combustion engine, a sequence of rotational speed values (φ.) is detected, the sequence of rotational speed values (φ.) being periodic Function is approximated with a model function, where parameters This model function can be determined, depending on the parameter parameters of combustion (KG) be determined. Method according to claim 1, wherein the model function is a Fourier series and the parameters of the model function are Fourier coefficients (A _k , B _k ), the Fourier series being represented by the polynomial (P)

(φ.) is the respective speed value, (A₀ / 2) a DC component, (f _z ) the ignition frequency of the internal combustion engine, (φ) of the respective rotational angle value associated with the rotational speed value (φ.) and (k) the engine order , Method according to claim 2, wherein the Fourier coefficients (A _k , B _k ) are determined by the fact that during the operation of the internal combustion engine a scalar product over selected lines of the pseudoinverse (PI) of a system of equations (GS) with the speed values (φ.) Of the sequence of Speed values (φ.) Is calculated, wherein the system of equations (GS) corresponds to a combination of the sequence of rotational speed values (φ.) With the polynomial (P). A method according to claim 3, wherein values associated with the Fourier coefficients (A _k , B _k ) are denoted generally by (F _n ), the Fourier coefficients (F _n ) being assigned to the linear model KG = K 1 · F 1 + ... + K n · F n + K n + 1 for the determination of combustion parameters (KG). The method of claim 1 to 4, wherein the Characteristic of combustion (KG) the angular position is at which 50% of the energy supplied to the combustion chamber implemented or the angular position is at which the cylinder pressure is the maximum, or the indicated mean pressure is.