Method for vibration damping of a drive train by means of an electric machine

Abstract

A method for vibration dampening of a drive train is disclosed. The method includes determining torsional parameters of the drive train via a state controller based on reconstructed torsional parameters of the at least one flywheel mass detected by at least one observer, determining the reconstructed torsional parameters based on disturbances in both (i) a load torque (M.sub.last) applied at an output of a torque transmission device and (ii) an induced torque (M.sub.ind) transmitted via the torque transmission device, determining a compensation torque via the state controller based on the reconstructed torsional parameters, and controlling an electric machine via the state controller based on the compensation torque to compensate torsional vibrations.

Claims

1. A method for vibration damping a drive train including an internal combustion engine, the method comprising: providing a state controller, at least one observer, and a torque transmission device arranged between a crankshaft and a transmission input shaft, the torque transmission device including at least one flywheel mass having a predetermined moment of inertia, determining torsional parameters of the drive train based on reconstructed torsional parameters of the at least one flywheel mass detected by the at least one observer, determining the reconstructed torsional parameters based on disturbances in both (i) a load torque (M.sub.last) applied at an output of the torque transmission device and (ii) an induced torque (M.sub.ind) transmitted via the torque transmission device, determining a compensation torque based on the reconstructed torsional parameters, and controlling an electric machine via the state controller based on the compensation torque to compensate torsional vibrations.

2. The method according to claim 1, further comprising modulating the compensation torque (M.sub.regler) upon an operating torque (M.sub.boost) of the electric machine.

3. The method of claim 1, wherein the at least one flywheel mass includes a plurality of flywheel masses.

4. The method according to claim 3, further comprising: processing at least one torsional parameter of the torsional parameters, wherein the at least one torsional parameter is at least one of: an angle of rotation (.sub.1, .sub.2, .sub.3) of at least one flywheel mass of the plurality of flywheel masses, an angular difference (.sub.12) between two flywheel masses of the plurality of flywheel masses, a torsional velocity (.sub.1, .sub.2, .sub.3) of at least one flywheel mass of the plurality of flywheel masses, or a difference of torsional velocities (.sub.12) between two flywheel masses of the plurality of flywheel masses.

5. The method according to claim 3, wherein the at least one observer is embodied in a linear fashion and a non-linear behavior of the torque transmission device is reconstructed via a non-linear estimation or a non-linearity is compensated by a decoupling of the disturbances.

6. The method according to claim 5, further comprising performing an estimation of non-linear torsional parameters of the torque transmission device via a neuro-fuzzy system using the torsional parameters of two flywheel masses of the plurality of flywheel masses.

7. The method according to claim 5, further comprising performing an estimation via a harmonically activated neural network based on the torsional parameters at the output or a rotor of the electric machine.

8. The method according to claim 5, the at least one observer further comprises a mean proportional-integral observer performing an estimation of the disturbances via the mean proportional-integral observer of two torsional parameters at an output side.

9. The method according to claim 1, further comprising determining at least one parameter relevant for torsional vibrations of the internal combustion engine induced torque (M.sub.ind) and using this as an input parameter for preliminary control of the induced torque (M.sub.ind) in the at least one observer.

10. A method for damping vibrations in a drive train comprising: providing: an electric machine; a torque transmission device comprising a flywheel mass with a moment of inertia; a state controller for controlling the electric machine to provide a compensation torque to compensate torsional vibrations; and, an observer for detecting a rotational velocity or a rotational angle of the drive train; arranging the torque transmission device between an input shaft of a transmission and a crankshaft of an internal combustion engine; reconstructing torsional parameters of the flywheel mass using: the rotational velocity or the rotational angle; a load torque at an output of the torque transmission device estimated by the observer; and, an induced torque from the internal combustion engine transmitted via the torque transmission device; and, using the torsional parameters as input parameters of the state controller.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention is explained in greater detail based on the exemplary embodiments shown in FIGS. 1 to 3. Here it shows:

(2) FIG. 1 a block diagram of a drive train with an active torsional vibration damping in various embodiments,

(3) FIG. 2 a block diagram to illustrate a reconstructed motor torque transmitted via a torque transmission device, and

(4) FIG. 3 a block diagram to illustrate a control and observation system of an active torsional vibration damping.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

(5) FIG. 1 shows a block diagram 1 with the drive train 2, shown schematically. The flywheel masses 3, 5, 9 form here an X-mass flywheel with three masses, with the flywheel mass 3 showing an torque of inertia J.sub.1 connected to the crankshaft of the internal combustion engine. The flywheel mass 5 with the torque of inertia J.sub.2 is elastically and torsionally coupled via the spring device 7 and the friction device 8 to the flywheel mass 3 and forms the torque transmission device 4 in the form of a double-mass flywheel. The torque transmission device 4 is coupled in a torque-proof fashion via the output 11, for example to a transmission input shaft of a transmission or the like, and transmits the output torque M.sub.ab to said transmission input shaft. A torsional coupling to the flywheel mass 9 is given by the driving wheels via the coupling 10 with the torque of inertia J.sub.3 between the transmission input shaft and the driving wheels. In the exemplary embodiments shown the flywheel mass 5 is rotationally coupled to the rotor of the electric machine 6 or is formed thereby. Together with the internal combustion engine the electric machine 6 forms the hybrid drive of the vehicle, perhaps starts the internal combustion engine, and during braking events of the vehicle recuperates mechanical energy in the form of electric energy. For this purpose the electric machine 6 is controlled via the operating torque M.sub.boost by an overall control unit of the vehicle. From the overall control unit, additionally the internal combustion engine is controlled and transfers, depending on its control, the motor torque M.sub.vm, subject to torsional vibrations, to the flywheel mass 3. By the mechanical control path with the respective system conditions and a perhaps non-linear behavior of the torque transmission device 4 the combination of the motor torque M.sub.vm and the electric machine torque M.sub.em (FIG. 3) of the electric machine 6 is transmitted as the output.

(6) For the active torsional vibration damping of the drive train 2 the state controller 12 determines the compensation torque M.sub.regler, which is modulated upon the operating torque M.sub.boost, for example is impressed thereon, and in the simplest case added thereto. To this extent, the active torsional vibration damping by the electric machine 6 is independent from the overall control of the vehicle and thus can essentially be provided as an additional module for all hybrid drive trains, perhaps even retrofitted.

(7) The state controller 12 is operated with deduced state parameters, such as detectable or reconstructed torsional parameters of the drive train 2, for example angles of rotation .sub.1, .sub.2, .sub.3, torsional velocities .sub.1, .sub.2, .sub.3, and the like of the flywheel masses 3, 5, 9, with the respective angular differences and differential torsional velocities being formed and evaluated thereby. These torsional parameters are generated by the observers 13, 14, 15, with the design, arrangement, and connection of the observers 13, 14, 15 being shown in the form of several embodiments. In a first exemplary embodiment at least the observer 13 serves to reconstruct the internal combustion engine and/or its reconstructed motor torque M.sub.vm,r and/or its torsional parameters. This reconstruction occurs here based on the torsional velocities cal of the flywheel mass 3, thus the crankshaft and the torsional parameters such as the torsional velocity .sub.2 of the flywheel mass 5, i.e. of the secondary side of the torque transmission device 4 and the corresponding angle of rotation .sub.2. Via the reconstructed motor torque M.sub.vm,r in the observer 14 the X-flywheel of the drive train 2 is determined, for example depending on its features, such as dynamic and the like, and the load torque M.sub.last is identified that is applied at the output, and the torsional parameters of the flywheel masses 3, 5 are reconstructed. The electric machine 6 is identified in the observer 15 via predetermined amperages I, the angle of rotation .sub.2, or the like, and the reconstructed electric machine torque M.sub.em,r is determined. Here the perhaps given non-linearity of the torque transmission device 4 is displayed in the observer 14. Alternatively the reconstructed motor torque M.sub.vm,r can be determined via the mean motor torque M.sub.vm,m, provided for example via CAN-bus, by way of estimation, a neuro-fuzzy system, a Kalmann filter, or the like.

(8) In a second embodiment, instead of the reconstructed motor torque M.sub.vkm,r, based on the same input parameters, the torque M.sub.ind induced at the flywheel mass 5 is estimated, which serves as the input parameter of the observer 14. In a third embodiment the induced torque M.sub.ind is determined via a preliminary control or a similar device from the position w.sub.L of the load lever, the upper dead center OT, and fed to the observer 14.

(9) In a fourth embodiment the observer 13 is omitted and the torques in the observer 14 considered as disturbances or decoupled in the form of induced torque M.sub.ind and the load torque M.sub.last, are identified via the torque parameters allocated to the torque of inertia J.sub.3, for example in the form of the angle of rotation .sub.3 and/or the torsional velocities .sub.3 and decoupled and/or estimated. For this purpose the observer 14 can be embodied as an unknown-input observer or PI observer, for example.

(10) With reference to the block diagram 1 of FIG. 1, FIG. 2 shows the block diagram 16, which illustrates schematically the identification of the torque transmission device 4 embodied as a double-mass flywheel and the reconstruction of the torque M.sub.ind induced on the output 11. The reference model 17 of the observer uses the angular difference .sub.12 and the different torsional velocities .sub.12 between the flywheel masses 3, 5. A linearization is yielded by using a set of local, linear modules 18, 19, . . . 20 in the reference model 17 characterizing the torque transmission device 4 according to the equation
M.sub.ind=c*.sub.12+d*.sub.12

(11) with the parameter factors c, d.

(12) Here weighing of the individual modules 18, . . . 20 occurs based on the weighing functions .sub.1, .sub.2, . . . .sub.M via the input parameters of the torsional velocity .sub.2 of the flywheel mass 5 and the angular difference .sub.12 of the two flywheel masses 3, 5. Alternatively, for example as a model, the following transmission function can be used of the torque transmission device
.sub.i(u)=w.sub.i,0+w.sub.i,1*u.sub.1+w.sub.i,2*u.sub.2+w.sub.i,3*u.sub.1.sup.2+w.sub.i,4*u.sub.2.sup.2+w.sub.i,5*u.sub.1*u.sub.2

(13) with the input parameters u.sub.1, u.sub.2 being used, for example the angular differences .sub.12 and the different torsional velocities .sub.12 (FIG. 2) and the parametering factors w.sub.i,0, w.sub.i,1, w.sub.i,2 . . . w.sub.i,5.

(14) FIG. 3 shows the block diagram 21, in which the observer 14a is embodied as a linear observer in different variants to identify the drive train, which perhaps behaves not linear due to the torque transmission device 4a, and to reconstruct their reconstructed torsional parameters. The inputs of the observer 14a include the block 23 from the operating torque M.sub.boost and the compensation torque M.sub.regler formed from the electric machine torque M.sub.em. In a first embodiment the neuro-fuzzy system 22 performs the non-linear estimate, processing as input parameters the torsional velocities .sub.1, .sup.2 of the flywheel masses of the torque transmission device, thus the primary and secondary side of a double-mass flywheel. The state controller 12a generates the compensation torque M.sub.regler from the torsional parameters of the observer 14a according to the state controller 12 and the observer 14 of FIG. 1. Here the motor torque M.sub.vkm transmitted via the torque transmission device 4a is identified as the interfering input and fed as the input parameter to the observer 14a. Here, an appropriate parametering of the state controller 12a allows a free input of system features, for example frequencies, damping factors, and the like of the torque transmission device. In another embodiment, using the linear observer 14a and waiving the neuro-fuzzy system 22, assuming a periodically occurring non-linearity in a predetermined torsional angle, a harmonically activated neural network can be trained. Here for each operating state the progression of the torsional parameters of the torque transmission device is determined preferably offline, i.e. for example before the use of the active torsional vibration damping, and for example saved as a table such that every operating state can be recalled in real time during operation and used for controlling the compensation torque. As an input parameter for the HANN here a torsional parameter at the secondary side is sufficient, for example the torsional velocity of a double-mass flywheel at the secondary side.

(15) In another embodiment, waiving the neuro-fuzzy system 22, the observer 14 can be provided with unknown inputs (unknown-input observer), for example the load torque M.sub.Last and the induced motor torque M.sub.ind. They are considered as unknown disturbances via methods of control technology. Two torsional parameters at the output side serve as unknown input parameters in the unknown-input observers to decouple the disturbances, for example torsional velocity at the secondary side and a wheel speed of the driving wheels. Here, for example the wheel velocity signals of the driving wheels can be used transmitted via the CAN-bus and subject to a reaction time.

(16) Accordingly, the linear observer 14a can be designed as a PI-observer according to an unknown-input observer. Unlike this one, in the PI-observer the disturbances are estimated as conditions instead of the disturbances being decoupling.

LIST OF REFERENCE CHARACTERS

(17) 1 Block diagram 2 Drive train 3 Flywheel mass 4 Torque transmission device 4a Torque transmission device 5 Flywheel mass 6 Electric machine 7 Spring device 8 Friction device 9 Flywheel mass 10 Coupling 11 Output 12 State controller 12a State controller 13 Observer 14 Observer 14a Observer 15 Observer 16 Block diagram 17 Reference model 18 Module 19 Module 20 Module 21 Block diagram 22 Neuro-fuzzy system 23 Block 24 Block 25 Block I Amperage J.sub.1 Torque of inertia J.sub.2 Torque of inertia J.sub.3 Torque of inertia M.sub.ab Induced motor torque M.sub.boost Operating torque M.sub.ind Induced torque M.sub.em Electric machine torque M.sub.em,r Constructed electric machine torque M.sub.last Load torque M.sub.regler Compensation torque M.sub.vm Motor torque M.sub.vm,r Reconstructed motor torque M.sub.vm,m Reconstructed motor torque OT Upper dead center w.sub.L Load lever setting .sub.12 Angular difference .sub.12 Difference of torsional velocity .sub.1 Weighing function .sub.2 Weighing function .sub.M Weighing function .sub.1 Torsional angle .sub.2 Torsional angle .sub.3 Torsional angle .sub.1 Torsional velocity .sub.2 Torsional velocity .sub.3 Torsional velocity