WO2008080380A1

WO2008080380A1 - Drive train

Info

Publication number: WO2008080380A1
Application number: PCT/DE2007/002247
Authority: WO
Inventors: Stephen John Jones; Bertrand Pennec; Andreas Walter; Thomas Winkler
Original assignee: Luk Lamellen Und Kupplungsbau Beteiligungs Kg
Priority date: 2007-01-05
Filing date: 2007-12-13
Publication date: 2008-07-10
Also published as: DE112007003032A5; DE112007003032B4

Abstract

The invention relates to a drive train, particularly for a motor vehicle having a drive unit controlled by means of a control device and having a drive shaft and an element, which is operatively connected to said drive shaft and is driven by the same, wherein at least one state value of the driven element is read into the control device, and a balancing of the cylinder of the drive unit is carried out utilizing the at least one state value.

Description

Antriebsstranq

The invention relates to a drive train and a method for operating the same in particular for a motor vehicle having a drive unit controlled by means of a control unit with a drive shaft and an element operatively connected thereto and driven by the latter.

Arrangements are known in a drive train in which an internal combustion engine is controlled by means of an engine management system. In particular actuators, such as injectors and injectors, and in spark-ignition engines, the spark plugs for determining the ignition timing are controlled by this. With the same control of the respective cylinders, which was in each case phase-shifted in the case of several cylinders of an engine, an "irregular running" of the drive unit due to combustion inequalities in the individual cylinders was nevertheless observed By evaluating an engine speed sensor, cylinder equalization has been sought by correcting, for example, ignition timing, injection quantities, and the like.

However, if driven on the crankshaft elements, such as ancillaries, clutch and / or transmission components such as dual-mass flywheels mounted that initiate dynamic torque dependent moments in the crankshaft, this may be due to a short-term misidentification of the crankshaft speeds due to this highly dynamic introduction of momentary torque changes to the Crankshaft result in a faulty interpretation of the currently applied engine torque, so that in the following, the engine management system controls the engine incorrectly. This can, for example, lead to over torques on driven elements and thus to overloading with premature failure of the latter, to loss of comfort and increased fuel consumption.

The object is therefore to propose a drive train and a method associated therewith for cylinder equalization of a drive unit which, despite time-varying and dynamic introduction of return moments of driven elements, permits improved equalization of the cylinders of the drive unit. Furthermore, there is the task to keep the cost of such a correction in terms of cost and simple and inexpensive. The object is achieved by a method for operating a drive train with a drive unit controlled by a control unit with a drive shaft and an element operatively connected thereto and driven by the latter, wherein at least one state value of the driven element is stored in the control unit and with the at least one Condition value, an induced engine torque is determined by means of which a cylinder equalization based on a Residuenbestimmung for individual cylinders of the drive unit takes place. Likewise, the solution of the task is carried out by the corresponding drive train using the designated method.

Under a control device is a computing unit with a microprocessor and a memory to understand, in the arithmetic unit, the necessary steps to control the drive unit are executed and recorded in a corresponding input unit corresponding sensor signals for characterizing the driver's desired torque and other input variables and with the aid of example in calculated engine data stored in an engine output map and output in an output unit to the units concerned, such as, depending on the output signal operated injectors or injection pumps. Other parameters, such as the engine speed or other present on the CAN bus data are usually also read to control the internal combustion engine. Usually, this control unit is connected to other control devices, for example for controlling the brakes, an optionally existing automated clutch or a control unit for controlling an optionally existing automated transmission or other control devices. It may also be advantageous to combine a plurality of control units to optimize the installation space and / or the electrical components or computing power of the control units to one or the number of total number of control units reduced number together and vote.

According to the invention, at least one state variable of a driven element is read into the control device for controlling the drive unit and optionally stored. The state quantity is then used in the algorithms for controlling the internal combustion engine, so that, using this at least one state variable, a behavior of the internal combustion engine adapted to the driven element is achieved. For example, can be carried out at least partially by the at least one state variable, the compensation of an interfering effect on the drive unit event resulting from the driven element. Under a drive unit in the context of this invention is an aggregate to understand that requires a control from the outside and having a drive shaft whose speed can be varied depending on a driver's request. Preferably, in this sense, a reciprocating piston engine with a crankshaft and also a rotary piston engine with a correspondingly designed drive shaft may be a drive unit. In a broader sense, the inventive embodiment applies to an electric motor with a drive shaft, which must be disturbed by a driven element and therefore must be readjusted by means of the state variable.

As a driven element, an element is understood, which receives a torque from the drive shaft or outputs a torque thereto. According to the invention this may be a dual-mass flywheel, which is connected on the one hand with a mass rotatably connected to the drive shaft and thus inelastically can absorb torque and on the other hand has an elastically coupled secondary mass, which in particular transmit a torque to the drive shaft in the dynamic operating condition and thus to a fault can lead the control of the drive unit. Furthermore, such components may be formed by other elements in the drive train, such as torque converters, transmission components such as shafts or gears or hybrid electric machines.

Driven elements may also be other components in the drive train in conjunction with the vehicle body or drive wheels, which lead in given circumstances to moments in the crankshaft and thus to disrupt the control of the internal combustion engine. For example, even a different adhesion of the wheels on a road to back moments lead. State variables applicable to such processes can be determined and also processed to compensate for disturbances in the control unit.

Furthermore, driven elements may be formed by ancillaries such as fuel pumps, power generators, air conditioning compressors and the like or belt drives, camshafts, adjusting devices and valve trains and the like.

In addition to the cylinder equalization, the following typical control tasks of the control unit, for example, by the drive train according to the invention and the associated method can be improved:

Motor control by means of engine torque, - A -

Signal, validation and adaptation of a real-time engine torque model, detection of combustion misfires present in the cylinder compensation control, idling control, anti-jerk control, load-stroke control,

Optimization of injection and ignition angle, optimization of injection quantity, detection and adaptation of pre-injection quantities, regulation and adaptation of engine start, regulation and adaptation of the starting process,

Determination of friction moments of a clutch between internal combustion engine and transmission,

Engine control unit performance check, general engine diagnostics, powertrain powertrain control, eg clutch, torque converter, all types of transmission, clutch slip control, slip control of drive wheels,

Control processes in hybrid applications, transmission protection,

Determination of the combustion chamber pressure of individual cylinders or determination of a dynamic mean pressure of the cylinders,

Analysis of components of drive units, such as dual-mass flywheel (position determination of components of a dual-mass flywheel).

It may be particularly advantageous if, in the presence of a plurality of state variables, these are combined in a so-called state space. In this case, a state space is to be understood as meaning a number of state variables which are each dependent on time and characterize a transmission system. With the aid of these state variables, the output variables necessary for controlling the transmission system can be determined or calculated in dependence on likewise time-dependent input variables in the assumption of known initial values.

In terms of physics, the state variables describe the energy content of a system, so that in addition to the equations for the dependence of the output variables on the state variables, the input variables and the initial values for the number of state variables As many differential equations can be formed, which reflect the dynamic behavior of the state variables.

From a vector perspective, in the case of linearity and time invariance of the system, the two systems of equations can be regarded as matrix systems whose individual coefficients can be represented and calculated in the elements of matrices. For a more detailed explanation of the treatment of state space models in measurement and control technology, reference is made to "Otto Föllinger, Control Engineering - Introduction to Methods and their Application, 8th Edition".

With the aid of the state, input, and output dependent data prepared in this way in the state space, a state space model can be developed that is identified and validated based on model data and / or empirically obtained data of the driven element. If state space model and these data match, a timely compensation of the engine torque of the drive unit, which is disturbed by the influence of the driven element, can be achieved at least partially.

Advantageously, the input variables used are data that are easily measurable. While measuring momentum for torque correction of the drive unit is comparatively complicated, it has been found that use of rotational speeds can be particularly advantageous. Thus, by inversion of the state space model from rotational speeds output quantities in the dimension of a moment can be obtained. In an advantageous manner, additional model data of the drive unit are recorded during the identification of the driven element, so that the corrected engine torque can be obtained directly as output variables. The corrected engine torque can be used in this way - as explained above with reference to typical application examples - large number of applications on the one hand in the operation of a vehicle with a drive unit and for the diagnosis of engine and / or powertrain behavior.

Particularly advantageous is the application of the proposed method for a drive train with a drive unit such as an internal combustion engine with a crankshaft and a dual mass flywheel mounted thereon with a primary part, which is accommodated on the crankshaft and a secondary part, wherein the two parts are supported and opposed the force of a spring device are limited relative to each other and with respect to the predetermined speed of the crankshaft rotatable relative to each other. The secondary part generally has a friction clutch, by means of which the secondary part and thus the drive unit can be coupled to the downstream in the direction of action transmission. It is understood that also modified embodiments in which an electrical machine is connected in parallel or in series, as well as all forms of power take-offs are covered by the invention. The dual-mass flywheel serves for vibration isolation of the oscillations generated by the combustion process of an internal combustion engine. In certain operating conditions, in particular during a relaxation of the compressed spring device, the dual mass flywheel is able to initiate moments of retraction on the internal combustion engine and to disturb by an unrecognized, short-term speed change the control behavior of the control unit for controlling the internal combustion engine, which can lead to said effects.

The dual-mass flywheel may comprise bow springs and / or short spiral springs as spring means for forming one or more damper stages, wherein the springs can also be combined by means of further force accumulators such as rubber elements and arranged on different diameters.

The spring device can be superimposed on a hysteresis device, which contributes in conjunction with the spring device for damping. The hysteresis device may be formed by two or more, in wet or dry rubbing contact with each other standing friction partners, which may be formed from axially or radially to each other, advantageously biased friction surfaces. Advantageously, a centrifugal force-dependent frictional contact during a relative rotation of the two parts, each having a predeterminable mass with an inertia thus forming upon rotation, be generated by the fact that over a circumference extending springs, such as bow springs, taken from a part, for example, the secondary part and at its outer radius under centrifugal force against a radial support of the other part, for example, the primary part, pressed and thus form a centrifugal force dependent hysteresis in a relative rotation of the two parts under which the two parts are difficult to rotate against each other. Further advantageous embodiments of friction devices may be, for example, against each other axial braced friction partners, of which the primary side and the other secondary side is arranged rotationally fixed or with torsional backlash. In terms of vibration, the dual-mass flywheel acts as a low-pass filter between the crankshaft and the transmission input shaft with a typical attenuation of the oscillation amplitude associated with a phase shift of the oscillation frequency. As already mentioned above, when the state space is transformed into a vector equation system, a system of linear equations must be present. In general, differential equations of system models, such as a dual mass flywheel, are not linear. However, if the controlled system has an approximate linear behavior in a certain working range, the differential equations can often be linearized for this sufficiently small selected working range. From this, linear state differential equations and output value equations can be established. The behavior of dual mass flywheels, in particular due to the bow spring characteristic with speed-dependent hysteresis, biased springs and clearance angle of the hysteresis, may be non-linear. According to the invention, therefore, the dual mass flywheel is linearized by the behavior of the dual mass flywheel is functionally represented by physical relationships and then differentiated to the operating points to be used. Particularly advantageous may be a further method of linearization, by means of which an approximated linear model is used and a range of validity is established in which an error occurring does not exceed a predetermined limit. The selection of such a model can be piece-specific. For example, in extreme cases, models may be designed as "black-box models" that very well approximate the input and output behavior and do not require any prior physical knowledge of the kinematic processes At the other extreme, the modeling of a state space can be such that the kinematics of the The subsequent identification then provides the desired compensation regardless of the model chosen, by choosing the variables of the state space such that there is an optimized match between the empirically or model-based data of a dual-mass flywheel and the input and output variables of the state space model is achieved.

In a particularly advantageous embodiment example, a simplified model for a dual-mass flywheel may be used, which includes an effective linear spring / damper element between two masses, namely a primary mass connected to the drive shaft and a secondary mass connected to the output. The state variables of this linear model agree with the nonlinear state variables of one sufficiently. It goes without saying that the choice of the model depends on the type of dual-mass flywheel and, depending on the design of the dual-mass flywheel - and in the wider sense also for each driven element - the corresponding models (black box models, gray box Models, white box models) can be optimized in an advantageous manner. For the simplified model, for example, the following relationships arise:

J _Pn ^■ co _p ή = M _pri - c ^■ (α -a _sec ) - d - (ω _ - ω _sec ) (1)

and

J sec ^k ω sec = ^{~ M} sec + ^C iCCpn ^{~ a} sec) + d ^• (ö _ - ω, _e4 ) (2).

In this case, J _pn and J ₅ denote the moments of inertia of the primary and secondary flywheels, M _pn and M _sec the primary and secondary mass of the dual mass flywheel, c the stiffness of the spring elements acting between the two masses and d the torsional damping of the damping elements acting between the two masses , a _pn and a _sέk the absolute _{angles of} the primary or secondary mass relative to the crankshaft and their derived from the time quantities of the angular velocities ω _pn and ω _sec and the angular acceleration ώ _pn and ώ _sec .

If the relative angle between the primary and secondary side and the rotational speeds of both flywheels are defined as state variables, the following state vector results:

Equation (1) and equation (3) result in the following state space representation:

The inventive idea relating to this exemplary embodiment provides only an evaluation based on the signals of the velocities ω _pri and ω _sec , so that the output vector y is given the following form:

It is understood that in the case of other parameters, such as the detection of accelerations, angles, and torques, the output vector can take different forms.

If a state space, as shown for example above, is created, the identification takes place for the driven element to be compensated, for example a dual-mass flywheel. According to the invention, a selection of physical input variables is determined for this purpose, which are sufficient for determining the estimated output variables. It has been found that a selection of irrelevant and redundant data leads to unnecessarily complex calculations and the lack of relevant input variables can lead to a lack of unambiguousness of the output variables.

It can be advantageous, for example, if data from a model calculation and / or from empirical tests are subjected to an outlier test, for example a median filter. Furthermore, it may be advantageous to filter the input variables, for example by means of a Butterworth low-pass filter. In such frequency-selective filters a cutoff frequency is specified up to which signals are transmitted almost unchanged. Higher-frequency signal components are strongly attenuated. It goes without saying that, in addition to or as an alternative, further advantageous filters can be used, adapted to an occurring noise in the input variables.

To further identify a driven element, such as here the dual mass flywheel, the order of the linear state space model must be determined. Advantageously, the order n = 3 for the elementary dual-mass flywheel, the ordinal number resulting from the number of state variables, proved by means of which an ideal spring / damper element can be described. The determination of the order must usually be made separately for each driven element and the underlying model.

After determining the system order, the usually continuous-time system must be converted into an equivalent discrete-time system. In other words, using the example of the dual-mass flywheel, a state space representation is present based on the data entered, which must then be converted into a time-discrete state space model. For this purpose, advantageously, a numerical evaluation of the general solutions of the state differential equation or a numerical integration of the state differential equation by means of numerical integration methods. To determine the discretized model and its variables, a so-called least-square estimator can then be used in which the distances between the output variables are set, preferably minimized, from a number of n measurements. The execution of the routine then leads to the parameters c, d, J _prj , J _S e _k defined in equation (4), that is, the dual-mass flywheel with its real behavior is mapped onto the model, it is identified. A subsequent validation of the underlying model can, for example, be carried out by calculating the model with given rotational irregularities and comparing it with a real behavior of a dual mass flywheel with the same rotational nonuniformities. At this point corrections can still be made to the model if there are major deviations.

A model of a driven element created in this way can now be stored in the control unit, and corresponding output variables can be introduced into the engine control. Using the example of the dual-mass flywheel, an estimate of the primary or secondary speed of the respective flywheels by specifying Motorbzw. Load moments take place. As a rule, torques in a motor vehicle can only be measured with a comparatively high outlay, so that according to the inventive idea an inversion of the state space model is advantageous. During inversion, the input variables of the inverted system then represent the rotational speeds of the primary and secondary flywheel mass, while the output variables represent torque values of the engine or load torque values. In this way, the induced motor torque can be determined by means of simple sensors, for example by means of speed sensors, to the primary and secondary flywheel. It is therefore possible to compensate the engine torque output as a setpoint with this induced engine torque in such a way that the influence of the dual mass flywheel on the engine can be at least partially eliminated. It is understood that such an inverted model should first be checked for invertibility and then re-validated. In this case, the load acting on the secondary side of the DMF by the drive train can be determined in an analogous manner, as a result of which an estimate of the load torque at the DMF can be achieved.

An advantageous embodiment of the invention is in addition to the compensation of the disturbed by the behavior of a dual mass flywheel engine torque a stationary reconstruction of the engine torque for cylinder equalization. For this purpose, it may be necessary to extend a so-called confidence interval, in which trusted data are determined and calculated, of the model by applying the dual-mass flywheel by means of so-called local linear models (LoLiMoT). The confidence interval describes the work area around the local identification work point, in which the model still provides sufficiently good results. By using local linear neuro-fuzzy models, complex modeling problems are broken down into many smaller and therefore simpler subsystems, which in turn can be described by linear submodels. For the application and implementation of the above mentioned filter systems as well as the LoLiMoT model, please refer to U. Kiencke / H. Jäkel "Signals and Systems", 3rd edition Verlag Oldenburg.

After the reconstruction of the engine torque (indicated engine torque), this invention can be used to calculate a working error or residual of individual cylinders. As a result, a torque value independent of the influences of the driven element or compensated for its influence is available. A determination of the residuals of each cylinder is done in Advantageously, from the torque balance of the engine, wherein the indicated in the manner described above engine torque is used.

For example, a typical operating error, a residual Rj for the cylinder i of a reciprocating engine can be derived from the general formulas for calculating the indicated _engine torque M _motor :

The indicated engine torque M _{Mot <} x, _r e _k is composed of the _gas torque M _gas , the mass torque M _Mass the valve torque Mv _e ntn and the friction torque M _Re jb together. Dividing the engine torque into its individual moments, the result is:

^{M '} Motor = ^M Gas ~ M ^' Mass ~ ^ Valve ~ ^M Rcib (6)

To calculate the work of the / -th cylinder, the engine torque is integrated via the cylinder-associated crankshaft segment. Ot ₁ corresponds to the center of the combustion process of the cylinder.

)there

In a four-cylinder engine, over 180 ° crankshaft angle (KWW) is exemplarily integrated, i. from top dead center (TDC) to bottom dead center (TDC), where the bottom dead center of the / -th cylinder corresponds to the top dead center of the / + 1-th cylinder. This results in:

-M _Mass -M _Valve -M _Friction ) da ₍₈₎

In a steady-state operating point, both the mass moment and the valve torque over 18O ⁰ KlAW are mean-free. The following applies:

and

The mass moment M _Mass (ω) is strongly speed-dependent. In high speed ranges (from approx. 3000 rpm), the mass moment even reaches the magnitude of the combustion torque M _Gas During an acceleration process, equation (9) is not valid due to the speed dependence of the mass moment. However, the average speed changes significantly within 180 ⁰ K-IzVWnJcM. At a medium speed of z. B. 800 rpm, the crankshaft rotates within 0.0375 seconds by 18O ⁰ KWW. In this small time window, the speed changes only insignificantly during an acceleration process. The integral

is therefore negligible even during an acceleration process relative to the integral of the combustion torque and the friction torque. This results in the integral of equation 8 in a stationary operating point

Equation 12 applies, as already described above, approximately also for acceleration processes. This results in the following relationship for the work: lV engine J _{or (/)} - Y ^v gas \ θT (i)

(13)

Simplified written gives you:

W ^vv engine, i - ~ W ^π gas, i - W ^rv Rc ibJ (14)

The combustion work of the cylinder i (Wβ _as , ι) can be divided into a reference work (WQ _aSιRef ) and a working _error the cylinder i (AW _{gas ι} ). The reference work is identical for each cylinder. This gives you:

W ₁ gas, i - ψ ^π gas, Re f + ^ ¹ A ^ W ^VV gas, i (15)

The sum of all work errors must be zero, since a previously set stationary operating point is not left by the occurrence of combustion inequalities. The condition therefore applies:

The average work of a cylinder, calculated from the work of a complete work cycle, is given by equations (14), (15), and (16):

W _motor , _M i _tt e _l = ₍₁₇₎

) (18)

In a stationary operating point, the average friction work results in:

l 1 ⁴

₄ . y ^ w Re ιb, i = W Re ib J (20)

/ = 1

Since the friction torque is generally speed-dependent, equation (20) does not apply during an acceleration process. However, the friction does not change significantly during acceleration in a working cycle [720 ⁰ KWW]. Therefore, equation (20) can be approximately assumed during an acceleration operation.

This results in the average work in a stationary operating point and approximately also in acceleration operations to:

γ ^γ engine, mean ~ ^γy gas .., Re f ^γγ frictionJ (21)

From the equations 14, 15 and 21 can now calculate the absolute work error

^ KW ^π engine, i = W ^vv engine, MMeI

- W ^VV gas, i - W ^γv Re ib, i -

- W ^vy Re ib, i))

^~ - W ^vr Gas, i - W ^vγ Gas, Re f

- ~~ W ^π gas, Re / + ^ ¹ ^ K W ^γγ gas, i - W ^π gas, Re /

= Δ W _{gas j} (22)

Replacing W _Motor ;, and W _M o _t o _r . _Ref by the corresponding integrals, one obtains for the absolute working error

π

^Δ W ^r motor, i = ^Δ W gas, i = \ ^{M '} motor ^{d (X} ~ ^ \ ^M motor ^da ₍₂₃₎ a. Π 0

2 If we now refer the absolute error of work to the reference work, one obtains the relative error, the so-called residuum

AW, gas, i

R, =

W ₁ (24)

Gas, re /

The reference work is calculated from the fuel energy supplied and thus results in too

^W Gas ^ f = n ^' QßMf = ⁷ I - ¹¹ U - ^m B, Ref • (25)

This results in the desired residual from the indexed engine torque M _Mo t _or :

Λ; - there (26)

Equation (26) holds for a stationary operating point. As the above considerations show, the residual computation according to equation (26) also provides sufficiently accurate results in acceleration operations.

By means of this calculation of the working errors of individual, several or all cylinders, the corresponding cylinders can be corrected, if necessary, in the next working cycle. In special cases and particularly designed engine concepts, it may also be advantageous to use a residuum obtained for a cylinder to improve the synchronization of the drive shaft also for the compensation of the firing of other cylinders, in particular if the cylinders are assigned in their firing order to one another such that a torque compensation the drive shaft can take place.

The invention will be explained in more detail with reference to Figures 1 to 3. Show this

FIG. 1 shows an inventive sequence of a control; Figure 2: an example of a dual mass flywheel explained creation of a

State space model

and

FIG. 3 shows a schematic embodiment of a procedure for cylinder equalization.

FIG. 1 shows an exemplary sequence for the control 1 according to the invention. In block 2, a setpoint for the control 1 is provided, which can be calculated as an initial value from a motor controller and output. In the case of an internal combustion engine, the desired value may be the desired torque which a driver prescribes by means of the accelerator pedal. Accordingly, this desired torque depending on other parameters, such as engine characteristics, gear engaged gear, operating condition of the vehicle, road conditions and the like can be adjusted. In block 3, the setpoint value 2 is applied to a controlled system, which may comprise the internal combustion engine and downstream parts such as, for example, a clutch, a dual mass flywheel or drive train components, such as ancillary units and subsequent transmissions. Using the example of a dual-mass flywheel, the target torque delivered to the internal combustion engine is disturbed by other components of the drive train through the two-mass flywheel by its torque-dependent rotation of the two masses against each other, so that in accordance with the invention, a state space model to compensate for the interference integrated to compensate for the disorder in block 5 becomes.

The creation of a state space model is explained in more detail in FIG. For operating the state space model, input variables are determined from the controlled system 3 and processed in the state space model. In accordance with the invention, input quantities can be all quantities that are available to the control unit as detection data, for example. For example, the control unit may be available on a so-called CAN bus, namely data such as speeds, accelerations, current and voltage values and the like. To compensate for the behavior of the dual mass flywheel, it may be particularly advantageous to resort to signals that detect the rotational speed of the primary and secondary mass, with their derived variables such as speed accelerations and angular velocities can be advantageous for evaluation. After passing through the state space model 5 output variables are output by this in the block 6, which can be read out and used via a suitable signal path 7 to correct the setpoint value 2. It is understood that such a control routine can be calculated serially or in parallel with other control routines in a control unit, for example by means of a microprocessor.

With reference to an example of a typical driven element in the form of a dual mass flywheel, the creation of a state space model 5 of Figure 1 in Figure 2 is explained in more detail.

As a basis for the mathematical and physical description of a dual-mass flywheel on the one hand, model data can be used in block 10, which are obtained from calculation programs, whereby corresponding simulation data of the real dual-mass flywheel are used as a basis. On the other hand, alternatively or additionally, experimental data may be provided in block 11 to a data set for the corresponding dual-mass flywheel shown in block 12, wherein these data can be obtained from test bench trials with a real dual-mass flywheel.

From the evaluation of the data in block 12, a state space model is developed in block 13 which maps the dual mass flywheel in the conditions specified for a state space. For example, linear equation systems must be available for the mathematical treatment of state space.

With the help of existing in block 12 and additionally present in block 15 data of an engine model, such as maps of an internal combustion engine, located in the state space 13 dual mass flywheel is identified, that is, in state, spatial coefficients are adjusted so that as possible great approximation between real measured data of the dual mass flywheel depending on fluctuating moments of the internal combustion engine and a real behavior of the dual mass flywheel result in the same engine conditions.

If the dual mass flywheel is identified, it is validated in the subsequent block 16. The validation determines the system accuracy of the state space and determines the boundary conditions of the applicability of the model. Motor data from block 15 can also be used for validation in block 16. In order to obtain output variables which correspond in terms of their dimension to the desired value and at the same time are based on simple input variables, it may be particularly advantageous to perform a so-called inversion in block 17. Originally, state space models are set up in energy units, whereby input variables in the dimension of moments are input and a corresponding behavior of the dual mass flywheel results in differential rotational speeds of the two flywheel masses. In order to avoid a cumbersome determination of measurable moments in a real environment such as in a motor vehicle, it is particularly in the case of determining disturbance variables of a dual mass flywheel advantageous to invert the state space modeled in block 13 in block 17. This means that the corresponding rotational speeds or their derived variables of the dual-mass flywheel are now used as input variables for the state space model as input variables, and thus motor torques or disturbance torques induced as output variables are available. It is understood that such an inverted model should also be checked for mathematical function (preferably stability) and validated.

The model available in block 18 after execution of the previous procedures is introduced as state space model 5 into corresponding control loops according to FIG.

FIG. 3 shows an exemplary embodiment of a cylinder-equalization procedure, as may be used in a controller for controlling a four-cylinder stroke piston engine. In a correspondingly adapted manner, the design example can be adapted to other engine types.

The cylinder equalization is described in detail on a cylinder (each with index 1). The remaining cylinders, r (indices 2, 3, 4) are analogously treated analogously or, if necessary, phase-shifted, so that a representation in dashed form is made only to complete them.

Starting with the driver's request torque Mwu _n sc _h , which transmits the driver via a load lever or an accelerator pedal to the control unit, this determines in node 20, for example from maps the individual moments for the individual cylinders, for example, injection quantities, ignition timing, valve closing angle and the like for the relevant Cylinder can be specified. The following node represents a miss that occurs at Treatment of the firing of the cylinder Z ₁ occurs. These may be faulty reactions of the components used, such as injection pumps, spark plugs, valves, etc., which may occur stochastically, for tolerance and / or life, and result in the cylinder Zi not being fired as calculated. As a result, the engine torque MM raw at the node 21 is erroneously formed from the individual moments Mi, M ₂ , M ₃ , M ₄ , each resulting from the firing of the individual cylinders. Of course, it can be assumed that the individual moments of the other cylinders different from Mi can be faulty in a similar manner.

The motor torque M _{M rOh} composed of these individual _moments is subsequently transmitted to the drive train with a driven element, here a dual-mass flywheel ZMS. To complete load moments acting on the dual mass flywheel M _load are shown. Due to a hysteresis of the dual mass flywheel - the secondary side is quasi wound at a moment thrust from the primary side and rotates slightly as the torque peak decays the primary side - the evaluation of the crankshaft speed n _pri for determining the crankshaft speed associated with a subsequent compensation of the errors in the cylinder firing also faulty. To improve this situation, therefore, the speed of the secondary side n _{SΘk is also} evaluated. This can be detected for example by means of a speed sensor directly to the secondary side of the dual mass flywheel or in the form of a sensor behind the usually provided on the dual mass flywheel clutch speed sensor, if the intermediate, the speed manipulating effects, such as clutch slip or translations considered or excluded become.

The two rotational speeds n _pr j, n _SΘk are input as input variables into the state space model and obtained according to the foregoing description, the induced reconstructed engine torque M _M otor. In a data collection, as well as the state. If the room model itself can be implemented in the control unit, the applied rotation values are recorded and checked for stationarity; if there is insufficient stationarity, a new data record is recorded. If the condition for stationarity is fulfilled, the residual calculation is performed. After passing through a control algorithm, the correction quantities K ₁ , K ₂ , ^" K ₃ , K ₄ into the nodal points-here by way of example the correction variable Ki into the node 20 for determining the firing conditions of the cylinder Z ₁ Return moments of a driven egg tents -showed on the basis of a dual-mass flywheel - by using a conditional space model.

LIST OF REFERENCES

Control routine Block for setpoint formation Block of a controlled system Block for reading input variables Block for the state space model Block for reading the output data Signal line Block for model data Block for experimental data Block for data record ZMS Block for state space Block for identification Block for data of the motor model Block for validation Block for inversion Block for model integration node node

Claims

claims

1. powertrain, in particular for a motor vehicle with a controlled by a control unit drive unit with a plurality of cylinders and a drive shaft and with this operatively connected and driven by this element, characterized in that at least one state value of the driven element is read in the control unit and with the at least one state value compensation of the cylinder is done among each other.

2. Drive train according to claim 1, characterized in that a plurality of state values are combined in a state space.

3. Drive train according to one of claims 1 or 2, characterized in that a einzuregelndes, acted upon by properties of the driven element with a disturbance torque induced motor torque of the drive unit is at least partially reconstructed by means of a determined from the at least one state value compensation value.

4. Drive train according to claim 3, characterized in that the reconstructed induced engine torque is determined timely.

5. Drive train according to one of claims 1 to 4, characterized in that the driven element is detected analytically by means of the at least one state value as a state space model.

6. Drive train according to claim 5, characterized in that in the state space model of the driven element, a model of the drive unit is integrated.

7. Drive train according to claim 6, characterized in that the analytical detection is made of data of a modeled driven element.

8. Drive train according to claim 5, characterized in that the analytical detection is carried out from experimentally determined data of the driven element.

9. Drive train according to one of claims 2 to 8, characterized in that in the state space, the movement of the driven element characterizing input variables entered and formed to determine the induced engine torque characteristic variables and output.

10. Drive train according to claim 9, characterized in that at least one input variable is a rotational speed.

11. Drive train according to one of claims 5 to 10, characterized in that the state space model is inverted.

12. Drive train according to one of claims 1 to 11, characterized in that the driven element is a mounted on the drive unit dual-mass flywheel with a drive shaft connected to the primary and a connectable to a downstream transmission means of a coupling secondary side.

13. Drive train according to one of claims 9 to 13, characterized in that at least the rotational speeds of the primary and secondary sides are used as input variables.

14. Drive train according to one of claims 9 to 13, characterized in that a non-linearized behavior of the dual mass flywheel is linearized.

15. Drive train according to claim 14, characterized in that the linearization takes place by differentiating non-linear equations of motion at a predetermined operating point.

16. Drive train according to one of claims 1 to 15, characterized in that for each cylinder determines a work error from the indicated moment and the corresponding cylinder is fired in a subsequent power stroke with the inclusion of the power stroke.

17. A method for controlling a drive unit with a plurality of cylinders in a drive train of a motor vehicle with a driven element characterized by the combination of the following features: at least one state quantity of the driven element is stored in a control unit; With the aid of the at least one state variable, an indexed engine torque which takes into account the reaction of the driven element to the drive unit is determined; the indexed engine torque determines a working error of the cylinders; The firing of the individual cylinders is corrected in the following working cycle by means of the working error determined for each cylinder.