DE102018210118A1 - Method for operating a fuel injection system of an internal combustion engine having a high-pressure accumulator - Google Patents

Abstract

The invention relates to a method for operating a fuel injection system of an internal combustion engine, which has a high pressure accumulator (115) and at least one injector (120), comprising a fuel line system (110) with at least one fuel pump (100) and at least one pressure sensor (140). for detecting the pressure in the high-pressure accumulator (115) and the fuel line system (110), in which it is provided in particular that a pressure value detected with the pressure sensor (140) is filtered by means of an extended Kalman filter (300, 305) and the pressure in the high-pressure accumulator (115) and in the fuel line system (110) is predicted on the basis of the pressure value filtered in this way.

Description

The invention relates to a method for operating a fuel injection system of an internal combustion engine having a high-pressure accumulator according to the preamble of claim 1. The present invention also relates to a computer program, a machine-readable data carrier for storing the computer program and an electronic control device by means of which the method according to the invention can be carried out is.

State of the art

A fuel injection system of an internal combustion engine affected here consists of a delivery unit (pump), its mechanical drive (e.g. via a camshaft or an electric drive), a fuel line system, a high-pressure fuel reservoir (rail) and several injection valves (injectors) ). The injection system is controlled by means of an engine control unit, the rail pressure being detected by means of a pressure sensor. For optimum control of the injectors, the rail pressure must be recorded as accurately as possible.

So it goes out DE 10 2008 043 130 A1 a fuel supply system for an internal combustion engine, in which the pressure prevailing in a fuel line is determined by means of a model-based estimation algorithm, in particular a state observer or a Kalman filter. This eliminates the need for a pressure sensor.

Furthermore, in DE 10 2007 019 640 A1 describes a method and a device by means of which the injection quantity in a fuel injection system can be metered precisely even without further technical measures. In this case, pressure values are corrected, which are provided with a time delay by a pressure sensor and are evaluated by a control unit.

Disclosure of the invention

The invention is based on the idea that the fluid pressure, for. B. in the high-pressure accumulator (rail), in a fuel line or on a high-pressure pump, not only to record under different operating conditions, but also to enable a precise prediction of the fuel pressure (i.e. prediction of the rail pressure).

As a result, it is possible to achieve an increased precision of the amount of fuel that is still to be injected or that has been injected after the injection has taken place, even at the beginning of a fuel injection. As a result, deviations in the metered fuel quantity relevant to emissions and requirements can be considerably minimized and, as a result, the requirement for subsequent exhaust gas purification processes can be reduced or even avoided altogether. This means that the stricter statutory emission regulations in many countries can be complied with at relatively low cost.

In the proposed method for operating a fuel injection system of an internal combustion engine having a high-pressure accumulator and at least one injector, which in addition to the high-pressure accumulator comprises a fuel line system with at least one fuel pump and at least one pressure sensor for detecting the pressure in the high-pressure accumulator and the fuel line system In particular, it is provided that a pressure value detected by the pressure sensor is filtered by means of an extended Kalman filter and the pressure in the high-pressure accumulator and in the fuel line system is predicted on the basis of the pressure value filtered in this way.

In comparison to known methods of pressure detection, the proposed method is based on the fact that an improved filtering of a measured pressure signal enables a more precise prediction of the pressure. The improved filtering is achieved by means of the extended Kalman filter (EKF).

In the case of a conventional Kalman filter, it is known that the most reliable value possible is ascertained on the basis of a large number of measurement values which are subject to uncertainties by means of computer-aided calculations. The processing speed depends, among other things, on the number of measured values supplied. By means of the proposed method, it is possible to only have to supply a Kalman filter with a smaller number of measured values.

With the proposed method, not only can noisy measurement signals be filtered very robustly without a time offset, but in particular the pressure curve in a high-pressure accumulator affected here can be precisely predicted. The pressure can be based on control signals for individual high-pressure components, e.g. Pumps or valves are predicted.

The minimum possible time difference of the prediction between the calculation of a control time (open time of an injection valve) in the control unit and the time of the actual injection of fuel corresponding to the fuel pressure corresponding to such a prognosis of the pressure curve is given by the already existing information defined in the control unit. At this point in time, the engine-side information for subsequent, regulated fuel delivery by means of the pump and the corresponding subsequent injection is already available.

The injected fuel mass can be better controlled by more precisely recording the pressure or the pressure curve. Downstream processes, e.g. Lambda control or exhaust gas aftertreatment can be considerably simplified. Further advantages result from a pressure control which is correspondingly optimized on the basis of a more precise pressure signal or a reduction in volume of the high-pressure accumulator which is made possible thereby and the pressure build-up times which can also be reduced as a result.

$${\underset{\_}{K}}_k={\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T\cdot {\left(\underset{\_}{H}\cdot {\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T+R\right)}^{-1}$$

$${\widetilde{\underset{\_}{x}}}_k={\underset{\_}{\widehat{x}}}_k+{\underset{\_}{K}}_k\cdot \left(\underset{\_}{y}-{\underset{\_}{\widehat{y}}}_k\right)$$

$${\underset{\_}{\tilde{P}}}_k=\left(\underset{\_}{I}-{\underset{\_}{K}}_k\cdot \underset{\_}{H}\right)\cdot {\underset{\_}{\widehat{P}}}_k$$

wobei

y_k:

Messgröße zum Zeitschritt k

x̂ _k:

geschätzter Zustandsvektor zum Zeitschritt k

x̃ _k:

korrigierter Zustandsvektor zum Zeitschritt k

h:

Funktion zur Schätzung der Messgröße

H:

Jacobi-Matrix der Messgrößenschätzfunktion h

K _k:

Kalman Verstärkungsmatrix zum Zeitschritt k

P̂ _k:

Kovarianz-Matrix zum Zeitschritt k

P̃ _k:

korrigierte Kovarianz-Matrix zum Zeitschritt k

R:

Messrauschen

I:

Einheitsmatrix

und wobei das Prädiktionselement durch die folgenden zwei Gleichungen (6) und (7) gebildet sein kann: $${\underset{\_}{\widehat{x}}}_{k+1}=\underset{\_}{f}\left({\underset{\_}{\tilde{x}}}_k,{\underset{\_}{u}}_k\right)$$

$${\underset{\_}{\widehat{P}}}_{k+1}=\underset{\_}{F}\cdot {\underset{\_}{\tilde{P}}}_k\cdot {\underset{\_}{F}}^T+Q$$

wobei

u _k:

Steuervektor zum Zeitschritt k

x̂ _k+1:

geschätzter Zustandsvektor zum Zeitschritt k+1

f:

Funktion zur Prädiktion des Zustandsvektors

F:

Jacobi-Matrix der Prädiktionsfunktion f

Q:

Systemrauschen.

In the proposed method it can be provided that the Kalman filter has a correction element and a prediction element, wherein the correction element can be formed by the following four equations (2) - (5): $$y_k=\underset{\_}{H}\left({\underset{\_}{\widehat{x}}}_k\right)$$

$${\underset{\_}{K}}_k={\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T\cdot {\left(\underset{\_}{H}\cdot {\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T+R\right)}^{-1}$$

$${\widetilde{\underset{\_}{x}}}_k={\underset{\_}{\widehat{x}}}_k+{\underset{\_}{K}}_k\cdot \left(\underset{\_}{y}-{\underset{\_}{\widehat{y}}}_k\right)$$

$${\underset{\_}{\tilde{P}}}_k=\left(\underset{\_}{I}-{\underset{\_}{K}}_k\cdot \underset{\_}{H}\right)\cdot {\underset{\_}{\widehat{P}}}_k$$

in which

y _k :

Measured variable for the time step k

x̂ _k :

estimated state vector for the time step k

x̃ _k :

corrected state vector for time step k

h :

Function for estimating the measured variable

H :

Jacobi matrix of the measured quantity estimator h

K _k :

Kalman time step gain matrix k

P̂ _k :

Covariance matrix for the time step k

P̃ _k :

corrected covariance matrix for time step k

R:

measurement noise

I :

identity matrix

and wherein the prediction element can be formed by the following two equations (6) and (7): $${\underset{\_}{\widehat{x}}}_{k+1}=\underset{\_}{f}\left({\underset{\_}{\tilde{x}}}_k.{\underset{\_}{u}}_k\right)$$

$${\underset{\_}{\widehat{P}}}_{k+1}=\underset{\_}{F}\cdot {\underset{\_}{\tilde{P}}}_k\cdot {\underset{\_}{F}}^T+Q$$

in which

u _k :

Control vector for time step k

x̂ _{k + 1} :

estimated state vector for the time step k + 1

f :

Function for predicting the state vector

F :

Jacobi matrix of the prediction function f

Q :

System noise.

In the proposed method, it can further be provided that the compression module of the respective fuel is additionally taken into account as an additional parameter in the expanded Kalman filter. As a result, the particular type of fuel present and / or the proportion of ethanol in the fuel and / or the content of water in the fuel can be determined.

In the proposed method it can also be provided that a Kalman state vector is expanded such that further parameters such as the quantity Qstat of the at least one injector and / or the delivery capacity of the fuel pump and / or the dynamics of a fuel line of the fuel line system are related Fuel pressure waves caused by fuel injection can be determined.

The likewise proposed electronic control device is set up in such a way that a fuel injection system of an internal combustion engine which has a high-pressure accumulator and at least one injector, which has a fuel line system with at least one fuel pump and at least one pressure sensor for detecting the pressure in the high-pressure accumulator and the fuel line system comprises, can be controlled by means of the proposed method.

The proposed electronic control unit has, in particular, an expanded Kalman filter, by means of which a pressure value detected by the pressure sensor is filtered, as a result of which the pressure in the high-pressure accumulator and in the fuel line system can be predicted on the basis of the pressure value filtered in this way.

The invention can be used in particular in a fuel injection system of an internal combustion engine, in particular of a motor vehicle, which has a high-pressure accumulator.

The computer program according to the invention is set up to carry out each step of the method, in particular if it runs on a computing device or a control device. It enables the method according to the invention to be implemented on an electronic control device without having to make structural changes to it. For this purpose, the machine-readable data carrier is provided, on which the computer program according to the invention is stored. By loading the computer program according to the invention on an electronic control device, the electronic control device according to the invention is obtained, which is set up to control an injection system or a corresponding internal combustion engine concerned by means of the method according to the invention.

Further advantages and refinements of the invention result from the description and the accompanying drawings.

It goes without saying that the features mentioned above and those yet to be explained below can be used not only in the respectively specified combination but also in other combinations or on their own without departing from the scope of the present invention.

list of figures

• 1 shows an affected fuel injection system of an internal combustion engine (in the present case four-cylinder) according to the prior art.
• 2 shows an exemplary pressure curve of a single injection with the indication of the average pressure from which the injection error can be determined.
• 3 shows a time sequence and calculations of a discrete extended Kalman filter.
• 4a shows a temporal course of a camshaft movement and supplied and discharged volume flows as well as a corresponding pressure course and measured values resulting thereby using a moving average filter and an EKF.
• 4b shows that in the top half of the 4a curves shown in an enlarged view.

Description of exemplary embodiments

This in 1 The injection system of an internal combustion engine shown comprises a delivery unit or pump 100 , whose mechanical drive 105 , e.g. B. a camshaft or electric drive, a fuel line system 110 , a high-pressure fuel reservoir or rail 115 as well as several injection valves or injectors 120 , These components are controlled via a (motor) control unit (not shown), with a first data line 125 the position of the mechanical drive, via a second data line 130 the pump 100 and over other data lines 135 the injectors 120 can be controlled. The hydraulic pressure p by means of an appropriate pressure sensor 140 detected.

• - Erfassen eines zuletzt gemessenen Druckwertes;
• - Erfassen eines Druckwertes mittels eines Moving-Average-Filters, z.B. bei einem 1 ms Raster über 10 ms mit einer Abtastrate von 100 pro s;
• - Durchführen einer linearen Interpolation der Druckwerte unmittelbar vor der Einspritzung;
• - Ausgleichen von systematischen Fehlern durch eine vorgebbare Regelmäßigkeit der Einspritzungen, indem vorausgegangenes Verhalten interpretiert wird, um den Druckwert in der Zukunft genauer abzuschätzen;
• - Erfassen eines Druckmittelwertes in einem definierten Messfenster (sog. kurbelwinkelsynchrone Messung);
• - Einsatz eines Drucksignalbeobachters zur sensorlosen Erfassung des Drucks.

In order to control the injectors precisely or optimally, the pressure must be recorded as accurately as possible p required in the rail. According to the current state of the art, this detection takes place in different ways, depending on the field of application. The following methods have become known:

• - Detection of a last measured pressure value;
• - Detection of a pressure value using a moving average filter, for example with a 1 ms raster over 10 ms with a sampling rate of 100 per s;
• - performing a linear interpolation of the pressure values immediately before the injection;
• - Compensation of systematic errors through a predeterminable regularity of the injections by interpreting previous behavior in order to estimate the pressure value more precisely in the future;
• - Detection of a mean pressure value in a defined measurement window (so-called crank angle-synchronous measurement);
• - Use of a pressure signal observer for sensorless detection of the pressure.

These methods, which are known per se, always interpret past events. It is therefore not possible with these methods to accurately predict the rail pressure p hold true.

If the rail pressure deviates, there is an error in the metered injection mass. This connection is in the 2 illustrated. The greater the deviation between a real, average pressure value p 215 and a pressure value pst present in the control unit and applied to an injector 205 is, the greater is the error resulting in the calculation of the actuation time (open time) of the injection valves. being the average pressure value p shown there 215 one during an open time 200 of the injector corresponds to the mean pressure. The dashed line 210 corresponds to an ideal rail pressure value.

The injected fuel mass error e can be calculated using the following equation (1): $$e=\sqrt{\frac{\overline{p}}{p_{St}}-1}$$

It can be seen from equation (1) that the error is very large, particularly at low pressure values (e.g. <10 bar in the case of an intake manifold injection). In the 4a and 4b are exemplary deviations in the fuel mass e shown, namely the fuel mass e _M (in the case of a direct petrol injection) in a conventional pressure detection and in a fuel mass determined according to the method described here e _K , It should be noted that in the present method the pressure value must be available in the control unit as information before the actual injection in order to be able to carry out the calculation of the activation time (open time) in good time. This time offset can further increase the error with regard to the metered injection mass.

The model-based rail pressure prediction described here is based in particular on a Kalman filter, which is known to enable measurements to be compared with model-based, predicted values. Assuming a linear model and white measurement and model noise, this approach leads to optimal results. In addition, such a Kalman filter approach can be implemented in a very resource-saving manner and thus without special additional measures on an electronic control unit (ECU) arranged in a motor vehicle.

For the present, non-linear high pressure system, however, the Kalman filter has to be modified to a so-called "Extended Kalman filter" (EKF). In the 3 is a sequential process of discrete EKF with division into correction 300 and in prediction 305 shown.

$${\underset{\_}{K}}_k={\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T\cdot {\left(\underset{\_}{H}\cdot {\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T+R\right)}^{-1}$$

$${\widetilde{\underset{\_}{x}}}_k={\underset{\_}{\widehat{x}}}_k+{\underset{\_}{K}}_k\cdot \left(\underset{\_}{y}-{\underset{\_}{\widehat{y}}}_k\right)$$

$${\underset{\_}{\tilde{P}}}_k=\left(\underset{\_}{I}-{\underset{\_}{K}}_k\cdot \underset{\_}{H}\right)\cdot {\underset{\_}{\widehat{P}}}_k$$

In the present exemplary embodiment, the correction is carried out using the following four equations (2) - (5), from which the meaning of the individual in 3 shown parameters: $${\widehat{\underset{\_}{y}}}_k=\underset{\_}{H}\left({\underset{\_}{\widehat{x}}}_k\right)$$

$${\underset{\_}{K}}_k={\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T\cdot {\left(\underset{\_}{H}\cdot {\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T+R\right)}^{-1}$$

$${\widetilde{\underset{\_}{x}}}_k={\underset{\_}{\widehat{x}}}_k+{\underset{\_}{K}}_k\cdot \left(\underset{\_}{y}-{\underset{\_}{\widehat{y}}}_k\right)$$

$${\underset{\_}{\tilde{P}}}_k=\left(\underset{\_}{I}-{\underset{\_}{K}}_k\cdot \underset{\_}{H}\right)\cdot {\underset{\_}{\widehat{P}}}_k$$

$${\underset{\_}{\widehat{P}}}_{k+1}=\underset{\_}{F}\cdot {\underset{\_}{\tilde{P}}}_k\cdot {\underset{\_}{F}}^T+\underset{\_}{Q}$$

In the present exemplary embodiment, the prediction is based on the following two equations (6) and (7): $${\underset{\_}{\widehat{x}}}_{k+1}=f\left({\underset{\_}{\tilde{x}}}_k.u_k\right)$$

$${\underset{\_}{\widehat{P}}}_{k+1}=\underset{\_}{F}\cdot {\underset{\_}{\tilde{P}}}_k\cdot {\underset{\_}{F}}^T+\underset{\_}{Q}$$

To illustrate the proposed method are in the 4a and 4b in the case of a direct petrol injection (BDE), the temporal movement curves of a camshaft of an internal combustion engine are shown. A known BDE system simulation using a simulation program "AMESim" and an extended Kalman filter using a second simulation program "MATLAB / Simulink" were carried out as a co-simulation. The 4a and 4b show the results of this simulation.

In the 4a as well as the in 4b The enlargement of the section shown is also illustrated for a camshaft movement, for example, resulting pressure measurements in unit [bar] over time in unit [s], which are supplied to the EKF. Values of the flow rates occurring at the four injectors in the present case are given on the ordinate in the unit [l / min]. It should be noted that from (only) in the 4a due to the period indicated by the horizontal arrow and starting from the vertically drawn dashed line, the EKF can no longer be supplied with pressure measurements.

Due to the camshaft speed, which is plotted in units [m / s] and varies periodically around zero 400 results in an in 1 shown pump 100 the illustrated course of the pump flow rate 405 . 405 ' . 405 ' in [l / min], essentially zero, with two brief increases 405 ' and 405 ' , The peak-like flow rate curves resulting from the camshaft movement 410 . 415 . 420 . 425 to the according 1 four injectors 120 are also shown in the unit [l / min].

In the top half of the 4a is also the rail pressure curve 430 shown in the unit [bar]. The rail pressure follows chronologically in Essentially the short increases in the pump flow rate 405 ' and 405 ' , Through the injections of the injectors, for example according to the flow rate curve 415 , there is a brief drop in the rail pressure 430 , The pressure signal provided to an EKF and shown in the unit [bar] 440 essentially follows the rail pressure curve 430 , In addition, the pressure signal curve provided to a moving average filter is 435 shown in the unit [bar].

From the in 4b The enlargement of the section shown can be seen more clearly, like the two quantities mentioned for the described equation (1), namely the fuel mass e _M in the case of direct petrol injection with conventional pressure detection and with a fuel mass determined according to the method described here e _K from the three curves 430 . 435 and 440 can be derived or calculated. This gives the fuel mass e _M as the difference between the actual rail pressure 430 and the pressure of the moving average filter at the time of the injection time calculation. The fuel mass also results e _K as the difference between that at the time of the injection time calculation or at the start of injection 440 predicted rail pressure and the actually prevailing rail pressure 430 ,

The equations and procedures used in the co-simulation are described below. The state space defined for this simulation example contains the following two states p and p : $$\underset{\_}{x}=\left[\begin{array}{c}p\\ \dot{p}\end{array}\right]$$

$$\underset{\_}{h}\left({\underset{\_}{\widehat{x}}}_k\right)=\left[p_k\right]$$

In contrast to the usual Kalman filter with a classic description in the state space, the functions f ( ^ x _k ; u _k ) from equation (6) and h ( ^ x _k ) from equation (2): $$\underset{\_}{f}\left({\underset{\_}{\widehat{x}}}_k.{\underset{\_}{u}}_k\right)=\left[\begin{array}{c}{\dot{p}}_k\cdot T_s+p_k\\ \frac{B}{V}\cdot {\displaystyle \Sigma Q}\end{array}\right]$$

$$\underset{\_}{H}\left({\underset{\_}{\widehat{x}}}_k\right)=\left[p_k\right]$$

The pressure p is described here with the first two terms of a corresponding Taylor series in combination with a discrete time step Ts. For the change in pressure over time p a hydraulic-capacitive expression is used. This expression is defined as the ratio of compression modulus (here B for bulk modulus) to hydraulic volume, multiplied by the sum of all occurring volume flows. The sum of these volume flows is made up of the incoming volume flow of the 1 shown pump 100 and the outgoing volume flow of the injectors 120 according to the following context ( 11 ) together: $${\displaystyle \Sigma Q=Q_p\cdot u_k\left(2\right)-Q_{Stat}\cdot \sqrt{\frac{p_k}{p_{Stat}}}\cdot u_k\left(3\right)}$$

$$H=\frac{\underset{\_}{h}\left({\underset{\_}{\widehat{x}}}_k\right)}{\partial {\underset{\_}{x}}_k}=\left[\begin{array}{cc}1& 0\end{array}\right]$$

For this purpose, in the present example implementation, three pieces of information from the control device are generated, according to the vector u _k , needed. The supplied volume flow of the pump is variable through the cam and is based on the input value Q _p = u _k (1) considered. The other control signals include the logical switching on and off of the components ( u _k (2) regarding the pump and u _k (3) for the injectors). Instead of a system matrix A and an output matrix C in the state space of a Kalman filter, the Jacobi matrices shown below become in the presently expanded Kalman filter F and H used: $$F=\frac{\partial \underset{\_}{f}\left({\underset{\_}{\tilde{x}}}_k.{\underset{\_}{u}}_k\right)}{\partial {\underset{\_}{x}}_k}=\left[\begin{array}{cc}1& T_S\\ \frac{B}{V}\cdot \frac{Q}{2\cdot \sqrt{p_{Stat}\cdot p_k}}\cdot u_k\left(3\right)& 0\end{array}\right]$$

$$H=\frac{\underset{\_}{H}\left({\underset{\_}{\widehat{x}}}_k\right)}{\partial {\underset{\_}{x}}_k}=\left[\begin{array}{cc}1& 0\end{array}\right]$$

In addition to the initializations of the Kalman state vector ^ x _k and the covariance matrix ^ P _k the covariance of the measured values is still parameterized R , the covariance matrix of system noise Q as well as the time constant T _s , These parameterizations are adjusted depending on the application. In the present example, only the prediction was included periodically ts called. The available measurement value is decisive for the correction loop, in this example the value 10 · T _s .

It should be noted that the described method can be expanded, for example with regard to the properties of fluids. In the present example, the compression module (B, bulk modulus) can also be taken into account as an additional parameter. Under the keyword "Flex-Fuels" it can be expected that modern PDC systems can be operated with all types of fuel. This results in a certain fluctuation range for parameter B. If this parameter is introduced as a state in equation (8) and the extended Kalman filter is adjusted accordingly, a parameter estimator for is also obtained size B. At the same time, the model becomes more precise on this point. This expansion of the state vector creates a variant of how the EKF can also be used. For example, the compression module B can be used to identify the fluid or the proportion of ethanol ( E5 . E10 etc.) as well as the water content in the fuel. An engine control unit can usefully use this information in other ways.

Also is the Kalman state vector ^ x _k principally expandable, so that further parameters like that Q _Stat of the injection valves or the delivery capacity (so-called "volumetric efficiency") of the delivery unit can be estimated. Furthermore, model extensions are conceivable, for example to map the dynamics of a line with regard to pressure waves.

The approach shown here can be used and implemented in any hydraulic system with an electrical control unit (e.g. injection system, power lift of a mobile machine or control of a hydraulic torque). In general, it is conceivable to use the method in all systems in which a physical pressure is formed by the inflow and / or outflow of a medium (fluid).

In the case of fuel injection, a measure for reducing emissions can be implemented right at the beginning of the entire functional chain, as a result of which subsequent exhaust gas aftertreatment devices can be subjected to less stress and optimized accordingly. It can be used in any injection system, in particular GDI ("Gasoline Direct Injection"), DI ("Diesel Injection"), PFI ("Port Fuel Injection"), P-DI ("Port and Direct Injection ”), WI (“ Water-Injection ”) or DWI (“ Direct-Water-Injection ”) are conceivable. The advantages of the described method come in particular with injection systems operating in the lower pressure range, e.g. WI and DWI, to wear.

The described method can be implemented in the form of a control program for an electronic control unit for controlling an internal combustion engine or in the form of one or more corresponding electronic control units (ECUs).

QUOTES INCLUDE IN THE DESCRIPTION

This list of documents listed by the applicant has been generated automatically and is only included for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 102008043130 A1 [0003]
• DE 102007019640 A1 [0004]

Claims (11)

Method for operating a fuel injection system of an internal combustion engine having a high-pressure accumulator (115) and at least one injector (120), which has a fuel line system (110) with at least one fuel pump (100) and at least one pressure sensor (140) for detecting the pressure in the high-pressure accumulator (115) and the fuel line system (110), characterized in that a pressure value detected with the pressure sensor (140) is filtered by means of an extended Kalman filter (300, 305) and the pressure in the high pressure accumulator (115) and in the fuel line system (110) is predicted. Procedure according to Claim 1 , characterized in that the Kalman filter has a correction element (300) and a prediction element (305). Procedure according to Claim 2 , characterized in that the correction element (300) is formed by the following four equations (2) - (5): $${\widehat{\underset{\_}{y}}}_k=\underset{\_}{H}\left({\underset{\_}{\widehat{x}}}_k\right)$$

$${\underset{\_}{K}}_k={\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T\cdot {\left(\underset{\_}{H}\cdot {\underset{\_}{\widehat{P}}}_k\cdot {\underset{\_}{H}}^T+R\right)}^{-1}$$

$${\widetilde{\underset{\_}{x}}}_k={\underset{\_}{\widehat{x}}}_k+{\underset{\_}{K}}_k\cdot \left(\underset{\_}{y}-{\underset{\_}{\widehat{y}}}_k\right)$$

$${\underset{\_}{\tilde{P}}}_k=\left(\underset{\_}{I}-{\underset{\_}{K}}_k\cdot \underset{\_}{H}\right)\cdot {\underset{\_}{\widehat{P}}}_k$$

Procedure according to Claim 2 or 3 , characterized in that the prediction element (305) is formed by the following two equations (6) and (7): $${\widehat{\underset{\_}{x}}}_{k+1}=f\left({\underset{\_}{\tilde{x}}}_k.{\underset{\_}{u}}_k\right)$$

$${\underset{\_}{\widehat{P}}}_{k+1}=\underset{\_}{F}\cdot {\widetilde{\underset{\_}{P}}}_k\cdot {\underset{\_}{F}}^T+Q$$

Method according to one of the preceding claims, characterized in that in the expanded Kalman filter (300, 305) the compression module of the respective fuel is additionally taken into account as an additional parameter. Procedure according to Claim 5 , characterized in that the particular type of fuel present and / or the proportion of ethanol in the fuel and / or the content of water in the fuel is determined. Method according to one of the preceding claims, characterized in that a Kalman state vector is expanded such that further parameters such as the size Qstat of the at least one injector (120) and / or the delivery capacity of the fuel pump (100) and / or the dynamics of a fuel line of the fuel line system (110) can be determined in relation to fuel pressure waves caused by fuel injection. Computer program which is set up each step of a method according to one of the Claims 1 to 7 perform. Machine-readable data carrier on which a computer program according to Claim 8 is saved. Electronic control device, which is set up, a fuel injection system of an internal combustion engine having a high-pressure accumulator (115) and at least one injector (120), which has a fuel line system (110) with at least one fuel pump (100) and at least one pressure sensor (140) Detecting the pressure in the high-pressure accumulator (115) and the fuel line system (110) comprises, by means of a method according to one of the Claims 1 to 7 to control. Electronic control unit after Claim 10 , characterized by an extended Kalman filter (300, 305), by means of which a pressure value detected with the pressure sensor (140) is filtered, whereby the pressure in the high pressure accumulator (115) and in the fuel line system (110 ) is predictable.

Images