DE102007012604B4 - Method for regulating injection of an injector of a direct injection internal combustion engine and direct injection internal combustion engine - Google Patents

Abstract

A method for controlling an injection of an injector (17) of a direct-injection internal combustion engine (10), wherein: - the control is based on at least one, at least partial cylinder pressure curve stored in relation to a crankshaft position, - the control has an at least partial injection course in at least one Cylinder (11), - upon completion of combustion of a first cycle and before commencement of injection of a subsequent cycle, a mathematical model description based on the stored cylinder pressure curve is formed, - the model description is used to form an injection profile for the subsequent cycle.

Description

The present invention relates to a method of controlling an injection of a direct injection internal combustion engine injector, a direct injection internal combustion engine, a vehicle, and a computer program.

The WO 2005/005813 A2 discloses a method for operating an internal combustion engine, in particular a diesel internal combustion engine with homogeneous fuel combustion. In this known method it is provided that a state variable in the cylinder is recorded as a function of the crank angle and a cylinder state signal is obtained therefrom, that at least two characteristic cycle characteristic values are determined from the cylinder state signal, that the determined cycle characteristic values with target values for the cycle characteristic values stored in a map are compared and an existing deviation between the two values is calculated, and that the deviation is fed to a control algorithm and the time of the fuel injection of at least one injection event and / or the proportion of inert gas in the cylinder are set as a manipulated variable in order to stabilize the combustion and / or Minimize noise and exhaust emissions. It is preferably provided that the 50% mass conversion point of the injected fuel and the maximum pressure increase in the cylinder are determined as cycle characteristic values. The pressure, the temperature, the ion current or the output signal of an optical measuring principle is preferably recorded as the state variable in the cylinder.

This known method is based on the consideration of dynamically calculating certain engine operating parameters, such as injection timing and exhaust gas recirculation rate, as a function of variables which describe the current state within the cylinder. To record the current cylinder state, for example, the pressure in the cylinder is detected as a function of the crank angle with a sensor. From this sensor signal, certain characteristic cycle parameters are subsequently calculated at an interval of 720 ° crank angle. The pressure curve within the cylinder is thus described by two characteristic values calculated from the pressure curve itself. In this known method, each of the currently determined cycle characteristic values is then compared with the desired value for the cycle characteristic values stored in a respective characteristic field as a function of engine speed and engine load, and an existing deviation between the two values is calculated. This deviation is subsequently fed to a control algorithm. The controller dynamically calculates the new engine operating parameters required to maintain the desired cylinder state, such as injection timing and recirculated exhaust mass.

The DE 197 49 817 A1 discloses an apparatus and a method for determining the start of injection or the combustion position in an internal combustion engine with at least one cylinder pressure sensor which generates a pressure proportional output signal, a crank angle sensor which outputs a signal representative of the crankshaft position signal, and an evaluation device, the pressure proportional output signal to the crank angle in Reference sets. In this known device and this known method it is provided that the measured pressure curve is compared in a predeterminable crank angle interval with at least one calculated taking into account thermodynamic contexts, stored in the evaluation pressure profile and that is recognized from the comparison result of the injection start or the combustion position, if the comparison result reaches a predefinable threshold. The evaluation device comprises a microprocessor which emits control signals, for example injection signals, to various components of the internal combustion engine as a function of the determined variables.

The DE 43 41 796 A1 discloses a method for controlling the combustion in the combustion chamber of an internal combustion engine. In this known method it is provided that in predetermined periods of time the pressure in the combustion chamber is detected, that the pressure values are stored during the compression stroke until a predefinable crank angle is reached and, after passing the predefinable crank angle, reproduced in mirror image form in equal time intervals, that the difference between the further recorded pressure values during combustion and the output pressure values during the compression stroke is formed, that the integral of the determined difference is formed, that a predeterminable surface portion of the integral determined and the associated crank angle, at which the predeterminable surface portion is reached, is determined that the deviation of the determined crank angle from a predefinable target crank angle is determined, and that the determined deviation represents a manipulated variable, which is supplied to a control unit for controlling the combustion.

In this document it is stated that the formation of the differential pressure integral calculates the area centroid, which should correlate very well with the true position of the combustion, and that based on this location of the centroid, a control of the ignition depending on the true position of the combustion and actual true Location of the combustion is possible. In the known method is provided that for the case of a deviation of the determined crank angle from the predefinable desired crank angle, a corresponding manipulated variable is determined based on the deviation, which then influences various control processes, such as ignition, lambda setting or exhaust gas recirculation. The predetermined desired crank angle is taken from a map. In the control unit, the determined manipulated variable, for example, is added to the characteristic field ignition angle and the ignition angle thus determined is output to the ignition output stage.

The DE 10 2004 001 118 A1 discloses a method for improved fuel metering, wherein a first variable based on a signal of a combustion chamber pressure sensor or a structure-borne sound sensor, which characterizes the combustion process in the combustion chamber of at least one cylinder, predetermined and at least one characteristic map of a second size, starting from the one manipulated variable for influencing the combustion process is predefinable, is read. The characteristic map and / or the second variable read out from the characteristic field are adapted as a function of at least one feature which is determined on the basis of the first variable.

The DE 101 59 017 A1 discloses a method for controlling an internal combustion engine, wherein at least one sensor detects a first variable that characterizes the pressure in the combustion chamber of at least one cylinder. Starting from this first variable, a second variable is determined which characterizes the change in the first variable and / or the course of the combustion and is dependent on the regulation of operating parameters of the internal combustion engine.

The EP 1 538 325 A1 discloses a system for controlling combustion in a diesel engine, the system including an injector, an exhaust gas recirculation system, a cylinder pressure sensor, a computer, and a controller. Based on the cylinder pressure measured by the cylinder pressure sensor and the crank angle position, the computer calculates at least one parameter characterizing the combustion, such as heat release, start of combustion or combustion phase. On the basis of the characterizing parameter, the control unit corrects at least one of the quantities injected fuel quantity, injection timing and amount of recirculated exhaust gas.

The object of the present invention is to simplify the regulation of an internal combustion engine.

This object is achieved according to the invention with a method according to claim 1, a direct injection internal combustion engine according to claim 33, a vehicle according to claim 43 and a computer program according to claim 45. Further embodiments are described in the subclaims.

The invention proposes a method for controlling an injection of an injector of a direct injection internal combustion engine. The control is based on an at least partial cylinder pressure curve, which is stored in relation to a crankshaft position. The regulation comprises an at least partial, preferably entire injection course in at least one cylinder, but may also comprise only a part thereof, in particular the part in which an injection course is initiated and / or carried out. Upon completion of combustion of a first cycle and prior to initiation of injection of a subsequent cycle, a mathematical model description is formed based on the stored cylinder pressure history. From the model description, an injection curve is formed for the subsequent cycle.

The mathematical model description, in contrast to the known methods of control using characteristic maps, describes the system to be controlled. The mathematical model can include, for example, one or more cylinders with an associated injector, the entire internal combustion engine as well as parts thereof, with the omission of associated units. However, add-on parts of other internal combustion engines can also be included, for example a compressor, in particular an exhaust gas turbocharger, and also a mechanical compressor, for example a roots charger, exhaust gas recirculation, exhaust gas aftertreatment or another device of the vehicle that influences the internal combustion engine. The mathematical model preferably includes functions. According to a first embodiment, at least one transfer function is used, for example, in order to determine the course of the injection for the next cycle. A further embodiment provides that, alternatively and / or additionally, equations, such as differential equations, are used. The mathematical model description or the mathematical model descriptions can be linear, non-linear, designed as neural networks and / or in some other way.

It can be provided that an actual injection profile for the cylinder is determined in a first cycle of a cylinder of the internal combustion engine. In the first cycle, an actual pressure curve for the cylinder is determined. In addition, at least one operating variable for the internal combustion engine is determined. The operating variables become a desired pressure curve for a later, second cycle formed of the cylinder. The mathematical model description is formed from the actual pressure profile and the actual injection profile. From the model description and the desired pressure curve, a desired injection curve is formed for the second cycle. In the second cycle, the desired injection curve is applied to the and / or at least one other cylinder of the internal combustion engine.

An advantage of this proposed method is that the combustion process can be improved by means of a pressure indication during driving, i. the actual pressure curve can be approximated very well to the desired pressure curve. The formation of the desired pressure profile can be effected, for example, by the various optimum desired pressure profiles being stored in the form of characteristic values, for example in the form of characteristic values of a cyclic process, and, if necessary, being calculated from these characteristic values; Thus, it is not necessary to store the complete target pressure profiles. In particular, according to a further development, the creation or determination of, in particular, a multiplicity of different, particularly complex characteristic diagrams can be dispensed with.

Preferably, the proposed method utilizes the idea that in each operating state there is a pressure profile that is optimal with regard to the respective specifications, such as emissions, fuel consumption, combustion noise, torque, wear, etc. Since many influences affect the mixture formation and combustion during operation, this optimum pressure profile can not or only rarely be achieved with an injection profile predetermined by a characteristic diagram, in particular amount and / or rate, as used in the known methods described above. The previous approach was to consider more and more of these influences in the regulation. This inevitably led to ever larger and / or more detailed and / or more numerous and / or memory-intensive maps and increasingly extensive calculations. Larger maps, however, require both a higher outlay in advance to generate them, in particular with regard to the application effort, as well as during operation in order to work with them, i. to read the desired values from them. In addition, larger data storage is needed. In contrast, according to the present invention, the system behavior is described with the help of the model description, so that many of the previously required maps can be omitted. Consequently, the memory requirement decreases and the application effort is essentially directed to the creation of optimal pressure profiles, for example, stored cylinder pressure profile and / or desired pressure profile.

Furthermore, the proposed method offers the advantages of an injection control or combustion process control, namely that cyclical fluctuations are damped and tolerances as well as aging and environmental influences, such as the height of the current location, outside temperature, air pressure, air humidity, oxygen, carbon dioxide, carbon monoxide and nitrogen oxide. , Fine dust content of the ambient air, etc., and other influences, such as fuel quality, can be largely compensated for.

Another advantage of the proposed method is that also operating modes are possible, which are not feasible with a conventional control / regulation, such as any shaping of the pressure curve, heating curve or combustion curve for conventional operation and / or regeneration operation, as well as a steady change between the progressions of a first mode and a second mode.

Although an internal combustion engine is a very complex and highly nonlinear system, if the pressure gradients and injection curves change only slightly from one cycle to a later cycle, which is the case in different operating ranges, it can be assumed to a good approximation System is linear and thus the model description formed in the one cycle will be used with good accuracy in the later cycle.

The method proposed in particular in the form of a transfer function uses, for example, an assumption that the effects of disturbance variables present during combustion, such as boost pressure, proportion of inert gas, etc., on the cylinder pressure curve or actual pressure curve are preferably already contained in this determined pressure curve and consequently automatically be taken into account in the model description. These effects therefore do not have to be specially modeled if it can be assumed that these disturbance variables are almost constant during a cycle. However, according to a further development, there is the possibility of being able to incorporate one or the other disturbance variable, such as, for example, speed, load, temperature, etc.

In the proposed method, the actual injection profile can be determined, for example, by measuring or, preferably, by reading it out of an individual injector map, which is usually already supplied with this injector, on the basis of a control signal for an injector that is to implement the injection profile for the cylinder becomes. The control signal for the injector can also be taken directly as the actual injection curve for the cylinder. The Model description preferably also the behavior of the "injector" system.

The term "measuring" should be understood here in a broad sense and includes above all a direct measurement and an indirect measurement, i.e. Deriving from at least one other measured variable. The actual pressure curve can also be determined, for example, by measuring. The operating variables can include the speed and the load, for example, and can be determined, for example, by measurement. The setpoint pressure curve can be formed, for example, by calculation or preferably by reading it from a map. This calculation of the target pressure curve can be done, for example, by storing the various optimal target pressure curves in the form of characteristic values, for example in the form of characteristic values of a cycle, in order to reduce the memory requirement and, if necessary, being calculated from these characteristic values; it is therefore not necessary to store the complete target pressure curves. In particular, according to a further development, it is possible to dispense with the creation or determination of characteristic diagrams for this. The model description and the target injection curve can be formed, for example, by calculation.

The distance between the first cycle and the second cycle can be arbitrarily selected as needed. For example, it may be provided that the first cycle and the second cycle follow one another directly or are separated from one another by at least one cycle.

Also, the time interval during which the actual injection curve or actual pressure profile are determined, can be selected arbitrarily as needed. Thus, for example, the determination of the actual injection profile and / or actual pressure profile can take place during at least one period of the cycle or during the entire cycle or during at least two cycles directly after one another.

Furthermore, the position of the time portion of the cycle can be arbitrarily selected as needed. For example, the period of the cycle may include the top dead center of the cylinder and / or the compression phase and / or the expansion phase of the cylinder.

It can be provided that the actual pressure profile and / or the actual injection profile is monitored for plausibility with the aid of at least one state variable of the internal combustion engine. As a state variable, for example, the exhaust gas temperature, the lambda value, the pressure in the intake manifold, the pressure in the exhaust manifold, etc. in question.

• - aus dem Ist-Einspritzverlauf bzw. Ist-Druckverlauf und wenigstens einem anderen Ist-Einspritzverlauf bzw. Ist-Druckverlauf von wenigstens einem anderen Zyklus und/oder von wenigstens einem anderen Zylinder ein gemittelter Einspritzverlauf bzw. ein gemittelter Druckverlauf gebildet wird;
• - die Modellbeschreibung unter Verwendung des gemittelten Einspritzverlaufs und/oder des gemittelten Druckverlaufs gebildet wird.

It can be provided that the model description is formed by:

• an average injection profile or an average pressure profile is formed from the actual injection profile or actual pressure profile and at least one other actual injection profile or actual pressure profile of at least one other cycle and / or at least one other cylinder;
• - The model description is formed using the averaged injection course and / or the averaged pressure course.

By means of this averaging, cyclic fluctuations of the profiles can be eliminated and measurement errors can be reduced.

• - aus dem Ist-Einspritzverlauf bzw. Ist-Druckverlauf durch Filtern, bevorzugt durch digitales Filtern, ein gefilterter Einspritzverlauf bzw. ein gefilterter Druckverlauf gebildet wird;
• - die Modellbeschreibung unter Verwendung des gefilterten Einspritzverlaufs und/oder des gefilterten Druckverlaufs gebildet wird.

It can be provided that the model description is formed by:

• - A filtered injection profile or a filtered pressure profile is formed from the actual injection profile or actual pressure profile by filtering, preferably by digital filtering;
• - The model description is formed using the filtered injection curve and / or the filtered pressure curve.

This filtering can reduce measurement noise.

The type of pressure curve can be chosen as required. Thus, for example, the actual pressure curve may be an actual cylinder pressure curve and / or the desired pressure curve may be a desired cylinder pressure curve, or the actual pressure curve may be an actual combustion pressure curve and / or the desired pressure curve may be a desired combustion pressure curve, or the actual pressure profile may be an actual heating profile and / or the desired pressure profile may be a desired heating profile, or the actual pressure profile may be an actual combustion profile and / or the desired pressure profile may be a nominal combustion profile.

• - aus dem Ist- bzw. Soll-Zylinderdruckverlauf zu Beginn der Kompression ein Ist- bzw. Soll-Schleppdruckverlauf gebildet wird;
• - der Ist- bzw. Soll-Schleppdruckverlauf von dem Ist- bzw. Soll-Zylinderdruckverlauf subtrahiert und derart der Ist- bzw. Soll-Verbrennungsdruckverlauf erhalten wird.

The actual or desired combustion pressure profile can thereby be formed from an actual or desired cylinder pressure profile such that:

• - From the actual or target cylinder pressure curve at the beginning of compression, an actual or target drag pressure curve is formed;
• - The actual or target drag pressure curve is subtracted from the actual or desired cylinder pressure profile and thus the actual or desired combustion pressure curve is obtained.

The calculation of the model description is made using this actual / target combustion pressure curve, actual / target heating curve or actual / target Burning process easier, since these courses only contain the information about the combustion and unwanted influences of the compression phase on the model description can be excluded.

The actual or target heating profile can be formed by calculation from an actual or target cylinder pressure profile, and the actual or target firing profile can be formed by integrating an actual or target heating profile.

The model description can be formed in different ways, for example with the aid of the least squares method or the method of auxiliary variables or the method of the stochastic approximation.

Likewise, the formation of the desired injection course can be carried out in different ways, for example according to a first variant by means of an inversion of the model description and / or according to a second variant by means of an inverse model description and / or according to a third variant by means of a universal inverse model description. In the first variant, it may be provided, for example, that initially a model description describing the pressure variation as a function of the course of injection or mapping the course of injection to the pressure profile is formed, then this model description is inverted, resulting in an inverse model description , which describes the course of injection as a function of the pressure curve or maps the pressure profile to the course of injection, is apparent, and then this inverse model description is applied to a, preferably predetermined, desired pressure profile, resulting in a desired injection profile. In the second variant, it may be provided, for example, that an inverse model description, which describes the course of injection as a function of the pressure curve or maps the pressure profile to the injection curve, directly, ie without the detour via a model description, the pressure profile is described as a function of the course of the injection or maps the course of the injection to the pressure curve, is formed, and then this inverse model description is applied to a, preferably predetermined, desired pressure profile, from which a desired injection profile emerges. In the third variant, it may be provided, for example, that a universal inverse model description that describes the course of injection in a wide range of application, ie not only in individual cases or for the respective current cycle, preferably generally valid, as a function of the pressure curve or the pressure curve depicts the course of injection, is applied to a, preferably predetermined, desired pressure profile, resulting in a desired injection curve. This universal inverse model description can for example be determined in advance by tests with engine test stands and / or test or pre-series vehicles from many injection curves and pressure curves and stored for the control, for example in a neural network, and then stored in operation with the current Measured values are adapted.

It may be provided that, for example, if the control does not converge according to the proposed methods, a different control method is used. However, this other control method can also be used otherwise, alternatively and / or supplementarily, simultaneously or in parallel with respect to time. This other control method may include, for example, a hitherto conventional control using maps and / or an adaptive control and / or a parametrically iteratively learning control. Such an inconvenience can occur, for example, in transient operation of the internal combustion engine, for example in the case of a sudden full load request or sudden overrun operation. The proposed method is preferably suitable for operation with diffusion-controlled combustion, so that if, for example, at low loads and thus small injection quantities only a premixed combustion is possible, then such an incongence can occur. Also, a change of the method may be provided as redundancy. A further development provides that with certain changes, in particular when exceeding at least one predefinable, the operation of the internal combustion engine characterizing value, in particular gradients, a changeover takes place with respect to the control of the internal combustion engine. If, for example, a load jump is detected, another method of regulation is used according to a development that does not use a model description or that provides for an extension. The extension can be done for example by an appropriately stored in maps adaptation as a function of load jump. According to one embodiment, the adaptation may take the form of a change in the model description or a supplement to the model description or an adaptation of the desired injection profile.

It can be provided that the regulation is combined with a precontrol according to this method. The precontrol can be made possible, for example, by a self-learning neural network, in which a large number of empirical values are available when driving through the same or similar operating situations. This can be used when an already known and / or similar operating situation is recognized Presetting of the already known and / or determinable values can be preset. In this way, a setting via the mathematical model description, in particular a transfer function, can also be applied to highly transient operating ranges, such as positive and negative acceleration phases.

Provision can furthermore be made for the formation of the model description to be supported by means of an implemented learning algorithm.

It can be provided that the model description comprises at least one transfer function. This can be linear or nonlinear. Two or more transfer functions can also be used. If there are several systems to be considered separately from each other, a mathematical model description, preferably in the form of a linear transfer function, can be present for each. For example, the internal combustion engine may be divided into various systems, such as a first and a second cylinder bank when a 6 or 12 cylinder engine is present. If a hybrid system is used, the first system may describe the internal combustion engine and a second system an electric motor and / or a generator.

• - der Soll-Einspritzverlauf des ersten Zyklus mit dem Soll-Einspritzverlauf eines früheren, dritten Zyklus verglichen wird;
• - der Soll-Einspritzverlauf des ersten Zyklus dann, wenn der Vergleich ergibt, dass dieser denjenigen des dritten Zyklus um mehr als einen erlaubten Änderungsbereich überschreitet, soweit begrenzt wird, dass er diese vorgegebene Grenze einhält.

It can be provided that the change in the injection curve is limited by:

• - The target injection course of the first cycle is compared with the target injection course of an earlier, third cycle;
• - The target injection course of the first cycle when the comparison shows that it exceeds that of the third cycle by more than an allowable range of change, to the extent that it is within this predetermined limit.

This limitation is particularly preferred when using linear transfer functions or linear equations, and allows a steady approach to the target profile to be achieved by repeated application of the above-proposed method using linear transfer functions or equations over several cycles can be, although the system itself, as already explained, is in principle non-linear. The desired injection course can also be generated as required in any desired manner, for example continuously, at a variable rate and / or by at least two separate injection events. The continuous generation can be realized, for example, with an injector, such as the CORA RS from FEV Motorentechnik GmbH, whose injection rate as well as injection times are extremely flexible and nevertheless precisely adjustable. More about a possible, preferably usable injector, its structure and its associated injection system are from the DE 100 01 828 A1 and the DE 10 2004 057 610 A1 to which reference is made in this disclosure in the context of this disclosure. The generation by at least two separate injection events can be realized with at least one conventional injector.

It can be provided that the model description is carried out with the aid of a neural network. With this neural network, a learning algorithm may preferably be implemented which supports the modeling of the model, for example by the formation of the transfer function and / or by the inversion of the transfer function and / or by the formation of the inverse transfer function.

The invention also proposes a direct injection internal combustion engine with a control device, at least one sensor and a stored functionality. The control device has at least one injector per cylinder, at least one load request determination and at least one calculation unit. With the help of the sensor, a course of an injection during a cycle of a cylinder can be determined. The stored functionality serves to form a mathematical model description on the basis of the injection course in the one cycle, which serves to form an injection course of a subsequent cycle.

The control device can be designed such that it carries out the control according to the method according to one of the preceding claims.

• - wenigstens einem Zylinder;
• - wenigstens einem Injektor je Zylinder;
• - einer Regelungsvorrichtung zum Regeln einer Einspritzung des Injektors;
• - einem Speicher, der ein Kennfeld mit Zylinderdruckverläufen für unterschiedliche Betriebszustände der Verbrennungskraftmaschine enthält;
• - einem Winkelsensor, der eine Winkelstellung einer Kurbelwelle der Verbrennungskraftmaschine erfasst;
• - wenigstens einem Drucksensor je Zylinder, der einen Druck in dem Zylinder erfasst;
• - einem Steuergerät, das mit dem Injektor, dem Speicher, dem Winkelsensor und dem Drucksensor verbunden ist; wobei das Steuergerät derart ausgebildet ist, dass es:
• - von dem Winkelsensor die Winkelwerte ausliest;
• - ein Ansteuersignal an den Injektor sendet;
• - aus dem Ansteuersignal und den Winkelwerten einen Ist-Einspritzverlauf als Funktion des Kurbelwinkels bildet;
• - von dem Drucksensor die Druckwerte ausliest;
• - aus den Druckwerten und den Winkelwerten einen Ist-Zylinderdruckverlauf als Funktion des Kurbelwinkels bildet;
• - nach Abschluss einer Verbrennung eines ersten Zyklus und vor Beginn einer Einspritzung eines späteren, zweiten Zyklus, aus dem Ist-Zylinderdruckverlauf des ersten Zyklus und dem Ist-Einspritzverlauf des ersten Zyklus die mathematische Modellbeschreibung bildet;
• - aus dem Kennfeld einen Soll-Zylinderdruckverlauf für den zweiten Zyklus ermittelt;
• - aus der Modellbeschreibung und dem Soll-Zylinderdruckverlauf für den zweiten Zyklus einen Soll-Einspritzverlauf für den zweiten Zyklus berechnet;
• - in dem zweiten Zyklus ein Ansteuersignal, das den Soll-Einspritzverlauf repräsentiert, an den Injektor sendet.

It can be provided that the internal combustion engine is provided with:

• - at least one cylinder;
• - at least one injector per cylinder;
• - A control device for controlling an injection of the injector;
• - A memory that contains a map with cylinder pressure profiles for different operating states of the internal combustion engine;
• - An angle sensor that detects an angular position of a crankshaft of the internal combustion engine;
• - At least one pressure sensor per cylinder, which detects a pressure in the cylinder;
• - A control unit, which is connected to the injector, the memory, the angle sensor and the pressure sensor; wherein the control device is designed such that it:
• - reads the angle values from the angle sensor;
• - sends a control signal to the injector;
• forms an actual injection curve as a function of the crank angle from the control signal and the angle values;
• - Reads the pressure values from the pressure sensor;
• forms an actual cylinder pressure curve as a function of the crank angle from the pressure values and the angle values;
• - after completion of combustion of a first cycle and before the start of an injection of a later, second cycle, the mathematical model description is formed from the actual cylinder pressure curve of the first cycle and the actual injection curve of the first cycle;
• - Determined a target cylinder pressure curve for the second cycle from the map;
• - Calculates a target injection profile for the second cycle from the model description and the target cylinder pressure profile for the second cycle;
• - In the second cycle sends a control signal, which represents the target injection course, to the injector.

• - wenigstens einem Zylinder;
• - wenigstens einem Injektor je Zylinder;
• - einer Regelungsvorrichtung zum Regeln einer Einspritzung des Injektors;
• - einem Speicher, der ein Kennfeld mit Zylinderdruckverläufen für unterschiedliche Betriebszustände der Verbrennungskraftmaschine enthält;
• - einem Winkelsensor, der eine Winkelstellung einer Kurbelwelle der Verbrennungskraftmaschine erfasst;
• - wenigstens einem Drucksensor je Zylinder, der einen Druck in dem Zylinder erfasst;
• - Mitteln, die mit dem Winkelsensor verbunden sind und die dazu dienen, dass sie von dem Winkelsensor die Winkelwerte auslesen;
• - Mitteln, die mit dem Injektor verbunden sind und die dazu dienen, dass sie ein Ansteuersignal an den Injektor senden;
• - Mitteln, die dazu dienen, dass sie aus dem Ansteuersignal und den Winkelwerten einen Ist-Einspritzverlauf als Funktion des Kurbelwinkels bilden;
• - Mitteln, die mit dem Drucksensor verbunden sind und die dazu dienen, dass sie von dem Drucksensor die Druckwerte auslesen;
• - Mitteln, die dazu dienen, dass sie aus den Druckwerten und den Winkelwerten einen Ist-Zylinderdruckverlauf als Funktion des Kurbelwinkels bilden;
• - Mitteln, die dazu dienen, dass sie nach Abschluss einer Verbrennung eines ersten Zyklus und vor Beginn einer Einspritzung eines späteren, zweiten Zyklus, aus dem Ist-Zylinderdruckverlauf des ersten Zyklus und dem Ist-Einspritzverlauf des ersten Zyklus die mathematische Modellbeschreibung bilden;
• - Mitteln, die mit dem Speicher verbunden sind und die dazu dienen, dass sie aus dem Kennfeld einen Soll-Zylinderdruckverlauf für den zweiten Zyklus ermitteln;
• - Mitteln, die dazu dienen, dass sie aus der Modellbeschreibung und dem Soll-Zylinderdruckverlauf für den zweiten Zyklus einen Soll-Einspritzverlauf für den zweiten Zyklus berechnen;

wobei:

• - die Mittel, die mit dem Injektor verbunden sind, des Weiteren dazu dienen, dass sie in dem zweiten Zyklus ein Ansteuersignal, das den Soll-Einspritzverlauf repräsentiert, an den Injektor senden.

It can also be provided that the internal combustion engine is provided with:

• - at least one cylinder;
• - at least one injector per cylinder;
• - A control device for controlling an injection of the injector;
• a memory containing a map with cylinder pressure curves for different operating states of the internal combustion engine;
• - An angle sensor which detects an angular position of a crankshaft of the internal combustion engine;
• at least one pressure sensor per cylinder, which detects a pressure in the cylinder;
• Means connected to the angle sensor and serving to read the angle values from the angle sensor;
• Means, which are connected to the injector and which serve to send a drive signal to the injector;
• - Means, which serve to form an actual injection curve as a function of the crank angle from the control signal and the angle values;
• - means, which are connected to the pressure sensor and serve to read from the pressure sensor, the pressure values;
• - Means, which serve to form an actual cylinder pressure curve as a function of the crank angle from the pressure values and the angle values;
• - Means, which serve to form the mathematical model description after completion of combustion of a first cycle and before commencement of injection of a later, second cycle, of the actual cylinder pressure curve of the first cycle and the actual injection curve of the first cycle;
• - means, which are connected to the memory and serve to determine from the map a target cylinder pressure curve for the second cycle;
• - Means, which serve to calculate from the model description and the target cylinder pressure curve for the second cycle a desired injection curve for the second cycle;

in which:

• - The means which are connected to the injector, further serve to send in the second cycle, a drive signal representing the desired injection course, to the injector.

The respective means can be integrated or spaced in integrated circuits, discrete circuits, in control units. For example, a motor controller, in which at least one of the means, in particular all means, is preferably integrated, can function as the sole control device for the regulation. However, there is also the possibility that there are different control units, each of which has different or the same means. The means can also be integrated in one or more computing and / or storage units, in particular using a CPU. Parallel operation can also be set up in order to effectively use different, similar control means, such as control devices. Such parallel operation is particularly advantageous in the case of a plurality of systems to be considered and in each case having mathematical model descriptions which are preferably also coupled to one another. The parallel operation allows a reaction of different systems, which is also to be made possible in particular in one work cycle, and which is coordinated with one another.

The pressure sensor can with a direct pressure measuring principle, as is the case for example in a piezoelectric pressure sensor, or with a working indirect pressure measuring principle, as is the case for example with an ion current sensor.

For the application and operation of this proposed direct injection internal combustion engine analogously apply the above statements to the proposed method.

At least one further sensor can be provided, which detects at least one state variable of the internal combustion engine. In this case it can further be provided that the control device is connected to the further sensor and is designed such that it reads out the values from the further sensor and with the aid of the state variables a plausibility diagnosis of the pressure sensor, preferably with the aid of on-board diagnostics , performs. Or means can be provided which are connected to the further sensor and which serve to read out the values from the further sensor and to carry out a plausibility diagnosis of the pressure sensor with the aid of the state variables, preferably with the aid of on-board diagnostics. The further sensor can be, for example, an exhaust gas temperature sensor, a lambda sensor, etc. Since the pressure sensor in the present invention is now a determining element for achieving an optimal burning process, the proper functioning of the pressure sensor can be monitored with the aid of the further sensor. In addition, these measures enable the pressure sensor signal to be corrected. It can be provided that the model description also includes a transfer function for this. This can be linear or non-linear.

It can further be provided that a neural network is provided, with the aid of which the functionality is implemented. This neural network can preferably be used to implement a learning algorithm which supports the formation of the model description, for example by forming the transfer function and / or by inverting the transfer function and / or by forming the inverse transfer function.

In the proposed method and the proposed internal combustion engine, the internal combustion engine can operate on a diesel principle and / or an Otto principle.

Preferably, the proposed method is used in a mobile device, in particular in a vehicle having an internal combustion engine according to the above proposals. The vehicle may be of any type and, for example, a passenger car, a truck, a rail vehicle, an airplane, a ship, etc. However, it may also be a stationary use, for example in an emergency generator, a combined heat and power plant or the like.

The invention further proposes a computer program for carrying out the proposed method. Preferably, the computer program comprises program code sections that cause the proposed method to be performed when the computer program is run on a computer.

• 1: eine schematische Darstellung einer ersten Ausgestaltung einer direkteinspritzenden Verbrennungskraftmaschine;
• 2: eine schematische Darstellung einer zweiten Ausgestaltung einer direkteinspritzenden Verbrennungskraftmaschine;
• 3: eine schematische Darstellung eines Kraftfahrzeuges;
• 4: ein schematisches Ablaufdiagramm einer ersten Ausführungsform eines Verfahrens zum Regeln einer Einspritzung eines Injektors der direkteinspritzenden Verbrennungskraftmaschine der 1;
• 5: ein schematisches Strukturdiagramm einer zweiten Ausführungsform eines Verfahrens zum Regeln einer Einspritzung eines Injektors der direkteinspritzenden Verbrennungskraftmaschine der 1.

Further advantageous embodiments and further developments are explained in more detail with reference to the following drawings. However, the resulting individual features are not limited to the individual embodiments. Rather, these features can be combined with further described above individual features or features of other embodiments to further embodiments. Show it:

• 1 a schematic representation of a first embodiment of a direct injection internal combustion engine;
• 2 a schematic representation of a second embodiment of a direct injection internal combustion engine;
• 3 : a schematic representation of a motor vehicle;
• 4 FIG. 4 is a schematic flowchart of a first embodiment of a method for controlling injection of a direct injection internal combustion engine injector. FIG 1 ;
• 5 FIG. 4 is a schematic structural diagram of a second embodiment of a method for controlling injection of a direct injection internal combustion engine injector. FIG 1 ,

The 1 shows a part of a diesel engine 10 in a first embodiment as an example of a direct-injection internal combustion engine. The diesel engine 10 has a cylinder 11 , a crankshaft 12 , which has a connecting rod with the piston of the cylinder 11 is connected, and a control device 13 for controlling an injection into the combustion chamber of the cylinder 11 on. An inlet pipe for fresh air and an outlet pipe 14 for exhaust gas flow into the combustion chamber and can be closed by means of an inlet valve and an exhaust valve.

The control device 13 has a control device in this first embodiment 15 and a memory connected to it 16 on a map with cylinder pressure curves for different operating conditions of the diesel engine 10 contains. The control unit 15 is also with an injector 17 that in the combustion chamber of the cylinder 11 leads to a pressure sensor 18 which has a pressure in the combustion chamber of the cylinder 11 detected, an angle sensor 19 , which is an angular position of the crankshaft 12 recorded, as well as an exhaust gas temperature sensor 20 and a lambda sensor 21 connected to the outlet pipe 14 are arranged. The control device 13 regulates the injection of the injector 17 depending on the signals from the pressure sensor 18 , the angle sensor 19 , of the exhaust gas temperature sensor 20 and the lambda sensor 21 , as exemplified below using the 4 will be explained in more detail.

In addition to this control procedure, the control unit performs 15 with the help of the exhaust gas temperature sensor 20 and the lambda sensor 21 , which the exhaust gas temperature and the lambda value as state variables of the diesel engine 10 continuously record a plausibility diagnosis of the pressure sensor 18 by.

The 2 shows part of a diesel engine 10 in a second embodiment, which is similar to the first embodiment. Therefore, only the differences between these two configurations will be described below. In this second embodiment, the control device 13 instead of the control unit 15 the first embodiment several means 22 to 31 , which are described in detail below. The means 22 are with the angle sensor 19 connected and serve to read the angle values from it. The means 23 are with the injector 17 connected and serve that they send the control signal to this. The means 24 serve to form the actual injection curve as a function of the crank angle from the control signal and the angle values. The means 25 are with the pressure sensor 18 connected and serve to read the pressure values from it. The means 26 serve to form the actual cylinder pressure curve as a function of the crank angle from the pressure values and the angle values. The means 27 serve to form the linear transfer function from the actual cylinder pressure curve and the actual injection curve after the combustion of the first cycle. The means 28 are with memory 16 connected and serve to determine the target cylinder pressure curve for the second cycle from the map. The means 29 are used to calculate a target injection profile for the second cycle from the transfer function and the target cylinder pressure profile for the second cycle. In this second embodiment, this is done similarly to the first embodiment, but it is provided here that the load sensor (not shown) is provided with the means 29 is connected and that they calculate the speed from the angle values. The means 23 also serve to send the control signal, which represents the desired injection profile, to the injector in the second cycle 17 send. In this second embodiment, the funds lead 22 to 29 the control procedure described above for the first embodiment.

The means 30 and 31 are with the exhaust gas temperature sensor 20 or the lambda sensor 21 connected and are used by the sensor assigned to them 20 . 21 Read out the values and with the help of these values the plausibility diagnosis of the pressure sensor 18 carry out. In this second embodiment, the funds lead 30 and 31 the plausibility diagnosis described above for the first embodiment.

The 3 schematically shows a passenger car 32 that with a diesel engine 10 is equipped according to the first or second embodiment. Other examples of such vehicles may be lorries, railcars, aircraft, ships, etc. Preferably, the method as well as the direct injection internal combustion engine is used when a biofuel and / or alcohol is added to a petroleum-based fuel.

The 4 shows a flow diagram of a first embodiment of a method for regulating an injection of the injector 17 of the diesel engine 10 of the 1 , In operation of the diesel engine 10 the crankshaft turns 12 , and the control unit 15 is designed such that it is from the angle sensor 19 continuously reads the angle values z, discrete or continuous. The control unit transmits at the beginning of a first cycle 15 in one step 101 a control signal to the injector 17 , which was determined either in the previous cycle or, for example, when starting the diesel engine 10 there is no previous cycle, it is specified and permanently saved. The control unit forms this control signal and the angle values 15 in one step 102 an actual injection curve EV ₁ (z) as a function of the crank angle z. In this first embodiment, the control unit reads 15 in the operation of the diesel engine 10 also from the pressure sensor 18 constantly the pressure values and forms in one step 103 from these pressure values and the angle values z an actual cylinder pressure curve p ₁ (z) as a function of the crank angle z.

After the completion of the combustion in this first cycle and before the start of an injection of the subsequent second cycle, the control unit forms in one step 104 from the actual cylinder pressure curve p ₁ (z) and the actual injection characteristic EV ₁ (z) is a linear transfer function G ₁ (z). The control unit can 15 recognize the corresponding point in time at which this formation of the transfer function starts, for example with the aid of a predetermined crank angle or by the fact that a current pressure value falls below a predetermined limit value.

Next, the controller accesses 15 on the memory 16 to and determined in one step 105 from the map stored there, a desired cylinder pressure curve p _soll (z) for the second cycle. This is done in this first embodiment in accordance with a load request, in one step 106 as one of the operating variables of the diesel engine 10 is determined. This determination of the load request is carried out here by means of a load sensor, not shown, with the control unit 15 is connected and detects the position of an accelerator pedal, for example, in a passenger car or truck. Another operating size for the diesel engine 10 , according to the specification of the target cylinder pressure profile is read from the map, here is the speed of the diesel engine 10 that the control unit 15 calculated from the angle values.

From the transfer function G ₁ (z) and _to this p target cylinder pressure curve (z) for the second cycle is calculated, the control device 15 in one step 107 then a target injection curve EV ₂ (z) for the second cycle. This happens here because the control unit 15 G inverts the transfer function ₁ (z) and then this inverse transfer function 1 / G ₁ (z) to the target cylinder pressure curve p _to (z) being applied. The control unit 15 converts in one step 108 then p _to this target injection characteristic (z) into a drive signal, and sends in a step 101 this to the injector in the second cycle 17 , The steps described above 101 - 108 are now repeated analogously for the second cycle. This cycle therefore represents a regulated route 109 represents.

The 5 shows the structure of a second embodiment of a method for controlling an injection of the injector 17 of the diesel engine 10 of the 1 , which is an extension of the first embodiment. Therefore, only the differences between these two embodiments will be described below. In this second embodiment, the controlled route 109 in series with a parametrically iterative learning control 110 , also abbreviated to ILR below. As a controller and controlled system of the regulated route 109 symbolic here are the control unit 15 and the injector 17 displayed. The output of the ILR 110 is with the entrance of the regulated route 109 connected, which is also an input of the regulator 15 is. The exit of the regulated route 109 , which is also the output of the controlled system 17 is, is with an input of the ILR 110 connected.

In this second embodiment, the control method of the first embodiment is combined with another control method, namely an ILR. For a more detailed explanation of the structure and functioning of an ILR, reference is made to the parallel patent application DE 10 2006 015 503 A1 referenced, the content of which is hereby fully referred to.

Claims (45)

Method for regulating an injection of an injector (17) of a direct-injection internal combustion engine (10), wherein: - The regulation is based on at least one, at least partial cylinder pressure curve, which is stored in relation to a crankshaft position; - The regulation comprises an at least partial injection course in at least one cylinder (11); after the completion of combustion of a first cycle and before the start of injection of a subsequent cycle, a mathematical model description is formed, which is based on the stored cylinder pressure curve; - An injection curve for the subsequent cycle is formed from the model description. Procedure according to Claim 1 , with the steps that: - an actual injection curve for the cylinder (11) is determined in a first cycle of a cylinder (11) of the internal combustion engine (10); - An actual pressure curve for the cylinder (11) is determined in the first cycle; - at least one operating variable for the internal combustion engine (10) is determined; - A target pressure curve for a subsequent, second cycle of the cylinder (11) is formed from the at least one operating variable; - The mathematical model description is formed from the actual pressure curve and the actual injection curve; a target injection profile for the second cycle is formed from the model description and the target pressure profile; - In the second cycle, the target injection curve is applied to and / or at least one other cylinder (11) of the internal combustion engine (10). Method according to one of the preceding claims, characterized in that the first cycle and the subsequent cycle follow one another directly or are separated from one another by at least one cycle. Method according to one of Claims 2 or 3 , characterized in that the determination of the actual injection course takes place during at least one period of the cycle or during the entire cycle or during at least two cycles directly following one another. Procedure according to one of the Claims 2 to 4 , characterized in that the determination of the actual pressure curve takes place during at least one time segment of the cycle or during the entire cycle or during at least two directly successive cycles. Method according to Claim 4 or 5 , characterized in that the period of the cycle includes the top dead center of the cylinder (11) and / or the compression phase of the cylinder (11) and / or the expansion phase of the cylinder (11). Procedure after a Claims 2 to 6 , characterized in that the actual pressure curve and / or the actual injection curve is monitored for plausibility with the aid of at least one state variable of the internal combustion engine (10). Method according to one Claims 2 to 7 , characterized in that the modeling of the model description takes place in that: - an average injection profile is formed from the actual injection profile and at least one other actual injection profile of at least one other cycle and / or of at least one other cylinder; - The model description is formed using the averaged injection curve. Method according to one of Claims 2 to 8th , characterized in that the modeling of the model description takes place in that: - an averaged pressure profile is formed from the actual pressure profile and at least one other actual pressure profile of at least one other cycle and / or of at least one other cylinder; - The model description is formed using the averaged pressure curve. Procedure according to one of the Claims 2 to 9 , characterized in that the model description is formed in that: - a filtered injection profile is formed from the actual injection profile by filtering, preferably by digital filtering; - The model description is formed using the filtered injection curve. Method according to one of Claims 2 to 10 , characterized in that the modeling of the model description takes place in that: - a filtered pressure profile is formed from the actual pressure profile by filtering, preferably by digital filtering; - The model description is formed using the filtered pressure curve. Method according to one of Claims 2 to 11 , characterized in that the actual pressure curve is an actual cylinder pressure curve and / or the desired pressure profile is a desired cylinder pressure curve. Method according to one of Claims 2 to 12 , characterized in that the actual pressure curve is an actual combustion pressure curve and / or the desired pressure curve is a desired combustion pressure curve. Procedure according to Claim 13 , characterized in that the actual combustion pressure curve is formed from an actual cylinder pressure curve in that: - an actual drag pressure curve is formed from the actual cylinder pressure curve at the beginning of the compression; - The actual drag pressure curve is subtracted from the actual cylinder pressure curve and the actual combustion pressure curve is obtained in this way. Procedure according to Claim 13 or 14 , characterized in that the target combustion pressure profile is formed from a target cylinder pressure profile in that: - a target drag pressure profile is formed from the target cylinder pressure profile at the beginning of the compression; - The target trailing pressure curve is subtracted from the target cylinder pressure curve and the target combustion pressure curve is obtained in this way. Procedure according to one of the Claims 2 to 15 , characterized in that the actual pressure curve is an actual heating curve and / or the target pressure curve is a target heating curve. Procedure according to Claim 16 , characterized in that the actual heating curve is formed by calculating from an actual cylinder pressure curve and / or the target heating curve is calculated from a target cylinder pressure curve. Method according to one of Claims 2 to 17 , characterized in that the actual pressure curve is an actual combustion curve and / or the desired pressure curve is a desired combustion curve. Method according to Claim 18 , characterized in that the actual combustion process is formed by integrating from an actual heating process and / or the desired combustion process by integrating a desired heating process. Method according to one of the preceding claims, characterized in that the modeling of the model is carried out using the least squares method and / or the auxiliary variable method and / or the stochastic approximation method. Method according to one of Claims 2 to 20 , characterized in that the formation of the desired injection course takes place with the aid of an inversion of the model description and / or with the aid of an inverse model description and / or with the aid of a universal inverse model description. Method according to one of the preceding claims, characterized in that if the regulation according to this method does not converge, another regulation method is used. Method according to one of the preceding claims, characterized in that the control is combined according to this method with at least one other control method. Method according to Claim 23 , characterized in that the control is combined according to this method with a map control. Procedure according to Claim 23 or 24 , characterized in that the control according to this method is combined with an adaptive control. Method according to one of Claims 23 to 25 , characterized in that the control according to this method is combined with a parametrically iteratively learning control. Method according to one of the preceding claims, characterized in that the control is combined according to this method with a pilot control. Method according to one of the preceding claims, characterized in that the formation of the model description is supported by means of an implemented learning algorithm. Method according to one of the preceding claims, characterized in that the model description comprises at least one transfer function. Procedure according to one of the Claims 2 to 29 , characterized in that the change in the injection course is limited in that: - the target injection course of the first cycle is compared with the target injection course of an earlier, third cycle; - The target injection course of the first cycle when the comparison shows that it exceeds that of the third cycle by more than an allowable range of change, to the extent that it is within this predetermined limit. Method according to one of Claims 2 to 30 , characterized in that the desired injection profile is generated continuously and / or by at least two separate injection events. Method according to one of the preceding claims, characterized in that the model description is carried out with the aid of a neural network. Direct injection internal combustion engine (10) with a control device (13), at least one sensor (18) and a stored functionality, wherein the control device (13) has at least one injector (17) per cylinder (11), a load request determination and at least one calculation unit (15). wherein the sensor (18) is arranged to determine a course of an injection during a cycle of a cylinder (11), and wherein the deposited functionality is designed to form a mathematical model description on the basis of the injection course in the one cycle which is designed to form an injection course of a subsequent cycle. Direct injection internal combustion engine (10) after Claim 33 , characterized in that the control device (13) is designed such that it controls according to the method according to one of Claims 1 to 32 performs. Direct-injection internal combustion engine (10) Claim 33 or 34 , with: - at least one cylinder (11); - At least one injector (17) per cylinder (11); - A control device (13) for controlling an injection of the injector (17); - A memory (16) containing a map with cylinder pressure profiles for different operating states of the internal combustion engine (10); - An angle sensor (19) which detects an angular position of a crankshaft (12) of the internal combustion engine (10); - At least one pressure sensor (18) per cylinder (11) which detects a pressure in the cylinder (11); - A control unit (15) which is connected to the injector (17), the memory (16), the angle sensor (19) and the pressure sensor (18); - The control device (15) is designed such that: - it reads the angle values from the angle sensor (19); - Sends a control signal to the injector (17); forms an actual injection curve as a function of the crank angle from the control signal and the angle values; - Reads the pressure values from the pressure sensor (18); forms an actual cylinder pressure curve as a function of the crank angle from the pressure values and the angle values; - after completion of combustion of a first cycle and before the start of an injection of a later, second cycle, the mathematical model description is formed from the actual cylinder pressure curve of the first cycle and the actual injection curve of the first cycle; - Determined a target cylinder pressure curve for the second cycle from the map; - Calculates a target injection profile for the second cycle from the model description and the target cylinder pressure profile for the second cycle; - In the second cycle, a control signal, which represents the desired injection profile, sends to the injector (17). Direct-injection internal combustion engine (10) Claim 35 , characterized in that: - a further sensor (20, 21) is provided which detects at least one state variable of the internal combustion engine (10); - The control device (15) is connected to the further sensor (20, 21) and is designed such that it reads out the values from the further sensor (20, 21) and, using the state variables, prefers a plausibility diagnosis of the pressure sensor (18) with the help of on-board diagnostics. Direct-injection internal combustion engine (10) Claim 33 or 34 , with: - at least one cylinder (11); - At least one injector (17) per cylinder (11); - A control device (13) for controlling an injection of the injector (17); - A memory (16) containing a map with cylinder pressure profiles for different operating states of the internal combustion engine (10); - An angle sensor (19) which detects an angular position of a crankshaft (12) of the internal combustion engine (10); - At least one pressure sensor (18) per cylinder (11) which detects a pressure in the cylinder (11); - Means (22) which are connected to the angle sensor (19) and which serve to read the angle values from the angle sensor (19); - Means (23) which are connected to the injector (17) and which serve to send a control signal to the injector (17); - Means (24) which serve to form an actual injection curve as a function of the crank angle from the control signal and the angle values; - Means (25) which are connected to the pressure sensor (18) and which serve to read the pressure values from the pressure sensor (18); - Means (26) which serve to form an actual cylinder pressure curve as a function of the crank angle from the pressure values and the angle values; - Means (27) which serve to form the mathematical model description from the actual cylinder pressure curve of the first cycle and the actual injection curve of the first cycle after completion of combustion of a first cycle and before the start of an injection of a later, second cycle ; - Means (28) which are connected to the memory (16) and which are used to determine a target cylinder pressure curve for the second cycle from the characteristic diagram; - Means (29) which serve to calculate a target injection profile for the second cycle from the model description and the target cylinder pressure profile for the second cycle; - Wherein: - The means (23), which are connected to the injector (17), further serve to send a control signal, which represents the target injection course, to the injector (17) in the second cycle. Internal combustion engine after Claim 37 , characterized in that: - a further sensor (20, 21) is provided which detects at least one state variable of the internal combustion engine (10); - Means (30, 31) are provided which are connected to the further sensor (20, 21) and which serve to read out the values from the further sensor (20, 21) and, with the aid of the state variables, a plausibility diagnosis of the pressure sensor (18), preferably with the help of on-board diagnostics. Direct injection internal combustion engine Claim 36 or 38 , characterized in that the further sensor is an exhaust gas temperature sensor (20) or a lambda probe (21). Direct injection internal combustion engine (10) according to one of Claims 33 to 39 , characterized in that the model description comprises at least one transfer function. Direct-injection internal combustion engine (10) according to one of the Claims 33 to 40 , characterized in that a neural network is provided, by means of which the functionality is implemented. Method according to one of Claims 1 to 32 or direct-injection internal combustion engine (10) according to one of Claims 33 to 41 , characterized in that the internal combustion engine (10) is designed to operate according to a diesel principle or def an Ottoprinzip. Vehicle (32), the direct injection internal combustion engine (10) according to one of Claims 33 to 42 having. Vehicle after Claim 43 which is a passenger car (32), a truck, a rail vehicle, an airplane or a ship. A computer program comprising program code sections that cause the method to be executed according to one of Claims 1 to 32 or 42 is performed when the computer program is running on a computer.

Images