DE102008012607A1 - Method for determining adaptation value for adjusting desired air-fuel ratio for fuel injection into internal combustion engine, involves predetermining desired value for air-fuel ratio of fuel injection for operating point - Google Patents

Abstract

The method involves predetermining a desired value for an air-fuel ratio of the fuel injection for an operating point. The lambda value of an exhaust gas of the combustion is recorded for observation of the air-fuel ratio. The adaptation value is applied corresponding to the positioning for correction of the pilot control of the fuel injection or air mass. Independent claims are included for the following: (1) a method for controlling a fuel injection of an internal combustion engine; and (2) a device for determining an adaptation value for an air-fuel ratio for a fuel injection into an internal combustion engine.

Description

The The invention relates to a method and a device for determining an adaptation value for the adjustment of an air-fuel ratio an injection system of an internal combustion engine according to the species the independent claims 1, 11 and 12. It is already known that in today's internal combustion engines in motor vehicles a load sensor is used with which sucked in the intake tract Air mass is determined. As a load sensor is usually an air mass meter for measuring the intake air flow and / or a pressure sensor for detecting an intake manifold pressure in the intake tract used. With the help of these sensors is the in a cylinder the internal combustion engine sucked air mass and determined tuned the required amount of fuel for one injected stoichiometric operation. The stoichiometric Operation is required to meet given legal requirements to reach the exhaust emissions. Furthermore, it is known that with the aid of the measured values of the sensors a comprehensive error diagnosis can be performed to possible causes to detect system errors.

Of the called load sensor is a not inconsiderable cost factor and thus the manufacturing cost of a motor vehicle considerably more expensive. Especially for smaller engines, for example be installed in low-cost motor vehicles and in particular are widespread in emerging markets, these are additional costs undesirable.

Of the Invention is based on the object, the determination of an adaptation value for setting a suitable air-fuel ratio an injection system of an internal combustion engine to simplify and thus make the injection system less expensive. This task comes with the distinguishing features of the sibling Claims 1, 11 and 12 solved.

at the method or the device according to the invention for determining an adaptation value for the setting an air-fuel ratio of an injection system with the features of the independent claims 1, 11 and 12 gives the advantage of being on a costly Load sensor can be omitted. In the inventive Procedure is therefore neither an air mass meter nor a pressure sensor required for measuring an intake manifold pressure. This can do that Injection system can be made much cheaper. In particular, motors with a low system complexity can be designed and manufactured much cheaper, because the load sensor by a stored model and a corresponding Adaption of learned correction values can be replaced. The cost-effective engines are in particular for Suitable vehicles, for example, in emerging markets are widespread. It is considered to be particularly advantageous that the exhaust emissions are about the same level of quality reach as with a load sensor equipped Injection system.

By those listed in the dependent claims Measures are advantageous developments and improvements that indicated in the independent claims 1, 11 and 12 Given method or the device. As a particularly advantageous is considered that with the inventive Method of adaptation value for the lambda controller only depending on the engine speed and / or on Based on a model determined engine load is determined. Thus, can in a simple way for each operating point of the internal combustion engine the appropriate amount of fuel to be injected can be calculated.

One Another aspect of the invention is that the adaptation value determined in a control unit as lambda control intervention becomes. As a result, the lambda controller is relieved, resulting in lower Controller strokes and thus for a faster compensation of mixture deviations leads.

In Another embodiment of the invention provides that the lambda control intervention is tuned to the cause of the error, either in the fuel path or in the air path can occur. In particular, it is provided that a relative error in the fuel path, for example at a improper slope of an injector is corrected.

at an offset error in the fuel path is provided that the lambda control intervention according to the formula FAC_LAM_COR = -100 · MFF_OFS / MFF_SP is determined. In the formula, the offset error of the fuel mass compared with a predetermined setpoint of the fuel mass. In the case of a deviation, a corresponding correction is carried out.

In a further embodiment of the invention, it is provided that a relative error in the air path can be corrected by a simple formula FAC_LAM_COR = â. For example, can be detected in this way, a faulty measured ambient pressure at a supercritical flow at the throttle and be corrected accordingly.

One Another aspect of the invention is that an offset error in the air path with the formula FAC_LAM_COR = 100 · MAF_OFS / MAF_SP is determined. In this formula, the offset error of the air mass flow compared with the predetermined setpoint of the air mass flow. at a deviation can thus be carried out a very simple correction become.

verwen det. Bei dieser Adaptionsfunktion treten drei Terme auf, mit denen ein Faktorfehler im Luftpfad/Kraftstoffpfad, ein Offset-Fehler im Kraftstoffpfad und ein Offsetfehler im Luftpfad erkannt werden kann. Bei iterativen Messungen kann mit Hilfe dieser Adaptionsfunktion eine Zuordnung des Adaptionswertes zu einer bestimmten Fehlerquelle erreicht werden. Durch die Unterscheidbarkeit verschiedener Fehlerursachen für die einzelnen Gemischfehler ergibt sich die Möglichkeit, dass eine gezielte Korrektur entweder im Luftpfad und/oder im Kraftstoffpfad durchgeführt werden kann.In a further embodiment of the invention, it is provided that the various causes for the mixture errors are determined as a function of the operating point with the aid of an algorithm. For this purpose, an adaptation function

used. In this adaptation function, there are three terms that can be used to detect a factor error in the air path / fuel path, an offset error in the fuel path, and an offset error in the air path. In iterative measurements, an adaptation of the adaptation value to a specific error source can be achieved with the aid of this adaptation function. The distinctness of different causes of error for the individual mixture errors results in the possibility that a targeted correction can be carried out either in the air path and / or in the fuel path.

According to the invention further provided that the algorithm for adaptation is performed iteratively, wherein in each adaptation step k is an updated value for one or more adaptation values w be determined.

From Another advantage is that for correcting the factor error f (N, MAF) for the proportion in the air or in the fuel path a division is performed. Thereby can advantageously, for example, aging influences individual components are taken into account.

One Embodiment of the invention is in the drawing and will become more apparent in the following description explained.

1 shows a schematic representation of an inventive device for determining an adaptation value for the adjustment of an air-fuel-air ratio,

2 shows a schematic structure of an inventive algorithm for adaptation and

3 shows a diagram with several learned adaptation curves.

Of the Invention is based on the consideration, a method or a device for an injection system of a motor vehicle in which the use of a load sensor in particular, for measuring the air mass flow or the intake manifold pressure at known injection systems is usual, be waived can. This allows the injection system to be much cheaper produced without relevant emission regulations get hurt. The inventive method is based essentially on an observation of the air-fuel ratio by means of a lambda probe whose measured values are compared with given model values depending on operating parameters the internal combustion engine are evaluated. As operating parameters a current speed N and a current load MAF is used, wherein the load MAF is taken from an adaptable model. The Deviations observed are due to an adaptation in the ongoing engine operation learned. Due to the structure of the deviations is trying to analyze if the cause of the deviation has occurred in the air path and / or in the fuel path. Based This assignment is determined iteratively adaptation values, which then for a correction of the pilot control of the injection system to be used. In this way, in every operating condition of the Combustion engine very precisely a stoichiometric air-fuel ratio be set. Thus, compliance with relevant emission regulations guaranteed even without the use of a load sensor.

in the The following will be explained in more detail initially, which error symptoms for stationary mixture errors can occur both in the fuel path and in the air path and what effects are expected.

The air mass flowing into an intake tract or into a cylinder of an internal combustion engine and the injected fuel quantity (fuel mass) are simulated by means of a model. With the help of the model, a predetermined setpoint for the air-fuel ratio (lambda value) is specified. Errors occurring in the modeling of the cylinder air mass and / or the introduced fuel mass are as Ab detected deviation from the predetermined setpoint value of the air-fuel ratio by a lambda sensor, which is arranged in the exhaust pipe in the vicinity of the cylinder outlet. This deviation is supplied to a lambda controller, which causes a corresponding correction of the amount of fuel to be injected. The correction is referred to as lambda control intervention FAC_LAM_COR and includes the sum of the relative errors in air mass and fuel mass. For determining the lambda control intervention FAC_LAM_COR, a nominal injector characteristic is specified. Due to the injector characteristic, an injected target fuel mass MFF_SP is assumed according to the formula as a function of the opening time of the injector TI: MFF_SP = c · Ti (1)

Therein, the constant c is a factor for the injector characteristic. This results in a realized injection time with a lambda control intervention TI_AV = MFF_SP / c · (1 + FAC_LAM_COR / 100) (2)

It The following types of errors in the fuel path can be distinguished become.

Relative error in the fuel path:

If the actually injected fuel mass MFF_AV is not equal to the target fuel mass MFF_SP, the relative error á (for example, with a wrong slope c of the injector characteristic) MFF_AV = (1 + á / 100) · c · Ti_AV (3)

If the error is fully corrected, then that actually corresponds injected fuel amount MFF_AV the predetermined target fuel amount MFF_SP. This formula is approximate for á << 100. Ti_AV is the actual opening duration of the fuel injector.

This results for the lambda control intervention FAC_LAM_COR = -á (4) Offset error in the fuel path

In the case of an offset error, with an extended or shortened fuel injector opening or closing time, an additional difference to the fuel mass MFF_OFS is calculated according to the formula MFF_AV = c * Ti_AV + MFF_OFS (5) injected. This results in the lambda control intervention FAC_LAM_COR = -100 · MFF_OFS / MFF_SP (6)

For Stoichiometric combustion is the target fuel quantity MFF_SP proportional to the cylinder air mass MAF_STK per cycle (stroke, stk) in (mg / stk) and thus the lambda control intervention FAC_LAM_COR is proportional to the offset error of the fuel mass MFF_OFS and indirectly proportional to the air mass MAF_STK in mg / stk.

Relative error in the air path

Analogous to the fuel path, errors also occur in the air path if the (actual) air mass flow MAF_KGH calculated from the model is not equal to the predetermined desired value MAF_SP. For example, a relative error â occurs when the ambient pressure at supercritical flow at the throttle is incorrectly measured. From the equation MAF_KGH = (1 + â / 100) · MAF_SP (7) results in the lambda control intervention FAC_LAM_COR = â (8) Offset error in the air path

An offset error occurs in the air path if, for example, the throttle valve is dirty. After the equation MAF_KGH = MAF_SP + MAF_OFS (9) results for the lambda control intervention FAC_LAM_COR = 100 · MAF_OFS / MAF_SP (10) It follows that the lambda control intervention FAC_LAM_COR is proportional to the error of the air mass offset MAF_OFS. In contrast to the fuel mass offset MFF_OFS, however, the resulting lambda correction is indirectly proportional to the air mass flow according to the equation MAF_SP = MAF_KGH - MAF_OFS (10a)

The cause relative errors in the air path and fuel path together to a constant relative lambda control intervention. A further distinction as to the cause of the error not possible. In contrast, the two have offset errors the air path or the fuel path has an effect on the lambda controller, Structurally, both from the factor errors and from each other different. This opens the way to a partial Distinguish the error when the system is in different operating states can be observed. Therefore, according to the invention, that several passes at different working points the internal combustion engine are driven, as later is explained in more detail.

Total mixture error

As above, results from the individual sources of error total mixture error. The total mixture error, its causes previously described, has an effect on the lambda controller. This distinguishes the system without load sensor to known Systems in which a load sensor is used. In the known Systems are largely compensated for errors from the air path the air path model with the load sensor adjusted so that - in Contrary to the device without load sensor - no influence on the lambda controller results.

In the case of the method or the device according to the invention, without the use of a load sensor, a lambda correction results for an arbitrary combination of the previously described error causes approximately according to the following formula:

Based An embodiment of the invention will be described below explains how it depends on the operating point of the internal combustion engine and the various mixture error causes built an algorithm for an adaptation function and can be used for an improved pilot control.

The structure of the adaptation function is selected analogously to equation (11) such that the dependency on the operating state shown above is maintained for the individual contributions. This results in the adaptation function:

The three terms on the right side of Equation 12 will follow explained in more detail.

The first term of Equation 12 relates to the factor error in the air / fuel path. The corrections of the factor error in the air and fuel path depend on the operating state of the internal combustion engine, in particular on the rotational speed N and the load MAF_STK. The speed N is detected by a speed sensor and the load MAF_STK can be taken from the stored model. For the two factor corrections, which can only be observed as a sum -á + â without a load sensor, a function f (N, MAF) is set, where N is the speed and MAF = MAF_STK. This function is subsequently to be adapted by Values w _i parameterized. The term STK means mass per stroke.

The second term of Equation 12 is characterized by the offset error in the fuel path. As already mentioned, the resulting factor correction is indirectly proportional to the mass air mass MAF_STK of a power stroke. The adaptation _value w _{MFF_OFS} is the associated proportionality constant. The adaptation _value w _{MFF_OFS} is thus proportional to the error of the fuel mass offset MFF_OFS.

The third term of Equation 12 corresponds to an offset error in the air path. As already mentioned, this mixture error is indirectly proportional to the desired air mass flow MAF_SP (in Kg / h) and the corresponding adaptation _value w _{MFF_OFS} is designated as an associated proportionality constant. The adaptation _value w _{MFF_OFS} thus corresponds to the offset of the air mass flow.

Realization in a (engine) control unit

The following is based on 1 a schematic representation of an inventive device for determining an adaptation value for the adjustment of an air-fuel ratio explained. Subsequently, a realization in the context of engine control based on the 2 ,

1 shows a schematic representation of an internal combustion engine 1 For example, a gasoline engine with a cylinder 5 in which a piston 4 is arranged by a connecting rod 3 is driven alternately while the piston 4 moved up or down. The combustion chamber of the cylinder 5 is via a suction tube 12 with an intake tract 10 or with an exhaust system 7 connected. In the intake tract 10 is first an air filter 15 arranged. The air filter 15 is a throttle 14 downstream, with the air flow L with a corresponding air mass within the intake 10 for example, directly or indirectly controlled by an accelerator pedal. Furthermore, the exhaust system 7 via an exhaust gas recirculation 8th and an EGR valve 9 with the suction pipe 12 connected. Inside the suction pipe 12 is the operating point-dependent intake manifold pressure pim available. Furthermore, an ambient pressure sensor 2 provided (AMP sensor), with which an ambient air pressure pamb is measured. Furthermore, on the suction pipe 12 an inlet 13 intended for crankcase ventilation. The combustion chamber of the cylinder 5 is opened or closed via an inlet valve E, so that via the inlet valve E, the cylinder 5 supplied fresh air can be controlled. Furthermore, an exhaust valve A is provided at the combustion chamber, with which the exhaust gas flow in the direction of an exhaust system 7 is controllable. Furthermore, a fuel injector 17 on the cylinder 5 (Cylinder head) arranged, with which the appropriate amount of fuel can be injected.

According to the invention it is further provided that at the outlet of the cylinder 5 in the area of the exhaust system 7 a lambda probe 21 is arranged, with which the residual oxygen is detected in the exhaust stream. The measured values of the lambda probe 21 are an indicator of the lambda value of the air-fuel mixture. The lambda probe 21 is with a motor control unit (programmable controller) 20 connected to the readings of the lambda probe 21 in conjunction with a lambda controller 22 prepared, as will be explained later. In the programmable controller 20 is a program stored with an algorithm, which corresponds to a current load of model values of the air path of the intake 10 a required fuel mass is calculated. This is the control unit 20 with the fuel injector 17 connected, which can be controlled accordingly. Furthermore, there is a memory 23 provided in the measurement data, models and programs with the algorithm (eg block 31 . 32 ) are stored. Furthermore, for the controller 20 an input for a speed N provided. It is envisaged that the algorithm according to the invention preferably in an already existing engine control unit 20 is implemented.

2 shows two different structures for determining an adaptation value for the adjustment of an air-fuel ratio. In the upper part of 2 (Block 31 ), the adaptation function is executed in a fast cycle, for example in the 10 ms cycle, as previously described in Equation 12. The adaptation function is guided by the lambda control intervention FAC_LAM_COR. The adaptation function is in a block 35 is modeled using an adaptive neural network NN and represents the function FAC_LAM_AD (N, MAF) represented by equation 12. The block 35 are input side according to the blocks 33 . 34 the current speed value and the model value for the air mass MAF supplied. The result is a lambda adaptation value FAC_LAM_AD, which is shown in block 36 is available as lambda correction value. The lambda correction value FAC_LAM_AD is fed to a unit LACO (corresponds to the lambda controller). The lambda controller forms the mixture control function for the fuel injection taking into account the lambda correction value FAC_LAM_AD, which is also referred to as the lambda adaptation value. The determined lambda adaptation value is transferred in the mixture control function LACO as additional multicorative correction for pilot control of the fuel mass to be injected.

The lower part of 2 shows a second cycle, which can be performed in a slower time frame, for example every 1000 ms. The lower part is as an adaptation part 32 educated. Here is the block first 39 fed to the mixture control function LACO. In the block 39 Thus, the lambda control intervention FAC_LAM_COR is included. Furthermore, in block 39 the lambda control intervention FAC_LAM_COR further observed with the aim that the lambda actual value corresponds to the lambda setpoint. The exit of the block 39 is further connected to the output of the block 36 with an adder 38 connected. Out of the block 38 result in the adaptation values for the current operating conditions of the internal combustion engine, which in block 37 weighted and learned with the help of the neural network. In the adaptation part 32 the weights of the adaptive neural network are always adapted to the effect that stationary no lambda control intervention FAC_LAM_COR is required. This means that in the ideal case the entire correction of the fuel mass to be injected is taken over by the neural network NN and thus the lambda controller based on the lambda signal is completely relieved. This makes it possible to considerably improve the emission behavior, since even in dynamic operation emissions-deviating deviations from the desired stoichiometric composition of the air-fuel mixture are prevented or at least considerably reduced.

adaptation algorithm

In a further embodiment of the invention, it is provided that the adaptation, ie the adaptation of the weights w _{MFF_OFS} , w _{MFF_OFS} preferably in the existing engine control unit Runaway leads. The algorithm is solved using an LMS algorithm (least mean squares). This algorithm is a real-time, iterative algorithm for solving a least squares regression problem. The solution method for the algorithm is off per se O. Nelles, loc. Cit., P. 62 as well as out Widrow & S. Steams, Adaptive Signal Processing, Prentice-Hall, London, 1985 known. For the adaptation several steps are needed. At each adaptation step k-1 to k, an updated value is calculated according to the rule for one or more adaptation values w _i w (K) i = w (K-1) i + η i · χ (K) i · e (K) e (K) = FAC_LAM_COR (K) (13) calculated. Here is χ (K) i the i-th regressor at the time k, which is calculated according to the rules to be suitably selected. The step sizes η _i determine the adaptation speed and are realized by suitable calibration variables to be selected. For the special case χ (K) i = 1 This procedure corresponds to the procedure already used today for lambda adaptation.

Adaptation of factor shares

in the Subsequently, the factor proportions of the adaptation values become corresponding the first term of the formula 12 described in more detail. The function f (N, MAF) should be represented by a small number of parameters and sufficient flexibility to represent different Own functions. Furthermore, the parameter adaptation in the Motor control stable and feasible with little effort be.

To solve this problem, in a further embodiment of the invention, a particularly suitable functional structure is proposed:
The entire load / speed range is divided into a predetermined number M of rectangular areas, with each of these areas being assigned an adaptation value w _i . In 3 an example of such a subdivision is shown. In 3 is a diagram in which on the X-axis, the speed N and on the Y-axis, the air mass per cycle MAF_CYL are plotted. The lower left rectangle in 3 corresponds to the rectangle no. 1, the right half of the large rectangle is the rectangle no. 2, the large upper left rectangle of 3 corresponds to the rectangle No. 3 and the three small left rectangles of 3 , which are adjacent to the rectangle number 1, correspond to the rectangles no. 4, 5 and 6.

Depending on the speed N and the load MAF_CYL, a series of curves with constant adaptation values are distributed over the rectangles. According to the theory, for a rectangle M, the adaptation value that runs approximately through its center is valid. The adaptation value changes in the direction of the edges, so that a smooth transition to the next rectangle M is given. For example, the adaptation value for the first rectangle is w ₁ = -16, for the second rectangle w ₂ = -2, for the third rectangle w ₃ = -7, for the fourth rectangle w ₄ = -15, for the fifth rectangle w ₅ = -15 and for the sixth rectangle w ₆ = -8. The relevant correction values are marked by circles in the individual rectangles.

Formal it is a neural network of the type local model network (LMN) with M = 6 local constant models.

According to the invention further provided that for the adaptation of the offset components for the fuel mass and the air mass accordingly the terms 2 and 3 in equation 12 also with the LMS algorithm is carried out as previously according to the equation 13 was explained.

• – wenn der Verbrennungsmotor warm ist und die Kühlwassertemperatur größer ist als ein vorgegebener Schwellwert,
• – wenn kein oder nur ein geringer Kraftstoffeintrag durch die Tankentlüftung entsteht,
• – bei einem stationären Betrieb des Verbrennungsmotors mit begrenzter Drehzahl- oder Lastveränderung,
• – wenn der Lambda-Regler aktiv ist bzw. bei einem System mit Lambda-Sprungsonde eine Adaption nur bei stöchiometrischem Betrieb erfolgt,
• – wenn keine Schubabschaltung aktiviert ist und
• – wenn der Regressor größer als der Schwellwert ist, der für jeden Regressor geeignet zu wählen ist.

The following activation conditions are provided for the adaptation in order to ensure the stability of the adaptation. The adaptation steps described above are preferably carried out only

• If the internal combustion engine is warm and the cooling water temperature is greater than a predefined threshold value,
• - if no or only a small amount of fuel enters through the tank ventilation,
• In a stationary operation of the internal combustion engine with a limited speed or load change,
• If the lambda controller is active or, in the case of a system with lambda jump probe, adaptation takes place only with stoichiometric operation,
• - if no fuel cut-off is activated and
• If the regressor is greater than the threshold that is appropriate for each regressor to choose.

Air and Fuel Path Correction Up to now it has been assumed that the adapted corrections are used exclusively for a correction of the fuel path. The introduction of the designations FAC_LAM_AD_COR for the correction in the fuel path and MAF_OFS for the correction in the air path thus far applied FAC_LAM_AD_COR = FAC_LAM_AD (N, MAF) MAF_OFS = 0 (14)

in this connection FAC_LAM_AD is calculated according to Equation 12.

With regard to the MAF offset component, however, it is advisable to correct the error at the location of its cause, ie in the air path. In extension of Equation 14, the following rule thus results for the calculation of a correction in the air and fuel path:

The additive correction in the air path MAF_OFS corrected according to equation 9, the air mass setpoint MAF_SP, so that the corrected air mass model value MAF_KGH yields.

Taking advantage of prior knowledge of typical tolerances in the air and fuel path, a further division of the learned corrections is advantageous. Thus, with a suitable calibration constant C_FAC_DISTR, it is possible to achieve an arbitrary distribution of the factor correction f (N, MAF) both on the air path and on the fuel path:

at Pollution or known aging behavior of the relative components C_FAC_DISTR can also depend on the mileage of the vehicle. For example, a faster aging in the injection system compared to the change the tolerances in the air path in the course of a vehicle life taken into account become.

In summary, the following advantages of the method according to the invention can be stated:
A distinction is made between different causes of mixture errors on the basis of different dependencies on the operating state.

The Adaptation of the individual tolerances takes place with a stable and efficient process.

The adaptation values can be used for the pilot control correction in the air and / or fuel path become.

Thereby a correct precontrol is possible even without a load sensor, so that no significant deterioration of emission behavior is to be expected.

QUOTES INCLUDE IN THE DESCRIPTION

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

Cited non-patent literature

• - O. Nelles, loc. Cit., Page 62 [0044]
• B. Widrow & S. Steams, Adaptive Signal Processing, Prentice-Hall, London, 1985. [0044]

Claims (12)

Method for determining an adaptation value for setting a desired air-fuel ratio for injection into an internal combustion engine ( 1 ), wherein for an operating point a target value for an air-fuel ratio of the injection is specified, wherein a lambda value of an exhaust gas of the combustion for monitoring the air-fuel ratio is detected and a lambda control ( 22 ) is used to correct the air-fuel ratio, characterized in that - during operation of the internal combustion engine ( 1 ) Measurements of the lambda probe ( 21 ) are used to determine a deviation from the predefined set value of the air-fuel ratio, - that the determined deviation is determined and learned as an adaptation value as a function of at least one operating parameter, - that the adaptation value is analyzed with respect to the cause and the air path and / or is assigned to the fuel path, and - that the adaptation value is used according to the assignment for correcting the precontrol of the injection and / or the air mass. A method according to claim 1, characterized in that the adaptation value in dependence on an engine speed (N) and / or an engine load (MAF) of the internal combustion engine ( 1 ) is determined. Method according to one of the preceding claims, characterized in that the adaptation value in a control unit ( 20 ) is determined as lambda control intervention (FAC_LAM_COR). Method according to one of the preceding claims, characterized in that for determining a relative error in the fuel path, a lambda control intervention according to the formula FAC_LAM_COR = -á where á is the relative error in the fuel path, for example, an inappropriate slope of an injector characteristic. Method according to one of the preceding claims, characterized in that for determining an offset error in the fuel path, a lambda control intervention according to the formula FAC_LAM_COR = -100 · MFF_OFS / MFF_SP is determined, wherein MFF_OFS / MFF_SP is the ratio of the offset error of the fuel mass to the predetermined target value of the fuel mass. Method according to one of the preceding claims, characterized in that for determining a relative error in the air path, a lambda control intervention according to the formula FAC_LAM_COR = â where? is the relative error in the air path, for example, a faulty ambient pressure at supercritical flow at the throttle. Method according to one of the preceding claims, characterized in that for determining an offset error in the air path, a lambda control intervention according to the formula FAC_LAM_COR = 100 · MAF_OFS / MAF_SP is determined, where MAF_OFS / MAF_SP is the ratio of the offset error of the air mass flow to the predetermined setpoint of the air mass flow. Method according to one of the preceding claims, characterized in that the following operating point-dependent adaptation function is formed for different mixture error causes

wherein the first term is a factor error in the air / fuel path, the second term is an offset error in the fuel path, and the third term is an offset error in the air path and where w _{MFF_OFS} , w _{MFF_OFS are} weighting factors for the adjustment of the fuel mass and the air mass, respectively. Method according to one of the preceding claims, characterized in that an adaptation value is calculated according to the rule for each adaptation step k - 1 → k: w (K) i = w (K-1) i + η i · χ (K) i · e (K) e (K) = FAC_LAM_COR (K) . where χ (k) / i is a suitable ith regressor at time k and η _{i is} the step size. Method according to one of the preceding claims, characterized in that for a correction of a factor error f (N, MAF) in the fuel and in the air path, a division according to the rule is carried out:

for the fuel path

for the air path, where C_FAC_DISTR is a calibration constant, for example an aging factor. Method for controlling an injection of a Internal combustion engine with a lambda control and with a fuel control, without a load measurement, wherein an adaptation value for the Lambda control and / or used for the fuel control which is determined according to one of claims 1 to 10. Device for determining an adaptation value for an air-fuel ratio of an injection system of an internal combustion engine ( 1 ) for a method according to one of the preceding claims, with a programmable controller ( 20 ), characterized in that a program of the control unit ( 20 ) is formed with an algorithm with which an adaptation value as a function of an operating point of the internal combustion engine ( 1 ) is determined and learned, that the adaptation value is assigned to the air path and / or the fuel path and that the adaptation value corresponding to the assignment to the pilot control of the injection system ( 17 ) and / or the air mass is used.