Device and method for reducing rotational imbalances of a drive train for a hybrid vehicle

Abstract

A system and a method for reducing a rotational imbalance of a drive train of a hybrid vehicle are provided. The drive train includes an internal-combustion engine, an electric machine, and a crankshaft. A reduction or nullification of the rotational imbalance takes place by actuating the electric machine. The actuation is implemented as an adaptive feed forward control, which provides an actuation signal for the electric machine. The actuation signal represents a desired torque to be generated by the electric machine, so that the electric machine outputs a torque that is at least approximately inverse with respect to the rotational imbalance to the crankshaft for the superimposition of the torque generated by the internal-combustion engine.

Claims

1. A system for reducing a rotational imbalance of a drive train of a hybrid vehicle, wherein the drive train includes an internal-combustion engine, an electric machine, and a crankshaft, the system comprising: an adaptive feed forward control that provides an actuation signal for actuating the electric machine, wherein: the actuation signal represents a desired torque to be generated by the electric machine, so that the electric machine outputs a torque that is at least approximately inverse with respect to the rotational imbalance, the adaptive feed forward control is configured to receive an estimate of an interference torque generated by an interference torque estimator based on a rotational speed of the electric machine, the adaptive feed forward control includes an interference variable monitor configured to receive a frequency vector from a frequency generator based on a rotational speed of the internal combustion engine, the frequency vector being in the form of a matrix A(ω,c), where ω represents a vector of individual frequencies corresponding to expected periodic oscillations associated with the rotational speed of the internal-combustion engine and c represents a vector of corresponding individual damping quantities, generate an expected oscillations vector in the form of x(k+1), where x(k+1)=A(ω, c)x(k)+Le(k), x(k) represents a previously determined expected oscillations vector and Le(k) represents a matrix of learning speeds for adaptation in response to the estimated interference torque, and generate the actuation signal for actuating the electric machine in the form of a desired electric machine output torque, the desired electric machine output torque being in the form of u=Cx(k), where C is a vector related to the frequency vector matrix A(ω, c) by the relationships of $A_i=\left[\begin{array}{cc}0& e^{-{\omega }_i{c}_i{T}_s}\\ -e^{-{\omega }_i{c}_i{T}_s}& 2e^{-{\omega }_i{c}_i{T}_s}\cos \left({\omega }_i\sqrt{1-c_i^2{T}_s}\right)\end{array}\right],C_i=\left[\begin{array}{cc}1& 0\end{array}\right]$ for an ith individual frequency ω.sub.i and ith individual damping c.sub.i in a scanning time T.sub.s between k and k+1 cycles of the adaptive feed forward control, and $A=\left[\begin{array}{ccc}{A}_1& \mathrm{.Math.}& 0\\ \mathrm{.Math.}& \ddots & \mathrm{.Math.}\\ 0& \mathrm{.Math.}& A_N\end{array}\right],C=\left[C_1\mathrm{.Math.}{C}_N\right],$ corresponding to superposition of N sinusoidal oscillations, and the adaptive feed forward control is configured to forward the actuation signal to the electric machine to as a control signal for generating and applying the desired electric machine output torque by the electric machine to the crankshaft for superimposition on a torque generated by the internal-combustion engine.

2. The system according to claim 1, wherein the frequency vector includes individual frequencies corresponding to expected periodic oscillations associated predefined engine orders.

3. A method for reducing a rotational imbalance of a drive train of a hybrid vehicle, wherein the drive train includes an internal-combustion engine, an electric machine, and a crankshaft, the method comprising: actuating the electric machine in response to an actuation signal provided by an adaptive feed forward control, wherein the actuation signal represents a desired torque to be generated by the electric machine, so that the electric machine outputs a torque that is at least approximately inverse with respect to the rotational imbalance, wherein the act of actuating the electric machine includes receiving by the adaptive feed forward control an estimate of an interference torque generated by an interference torque estimator based on a rotational speed of the electric machine, generating with a frequency generator of the adaptive feed forward control a frequency vector from a frequency generator based on a rotational speed of the internal combustion engine, the frequency vector being in the form of a matrix A(ω, c), where ω represents a vector of individual frequencies corresponding to expected periodic oscillations associated with the rotational speed of the internal-combustion engine and c represents a vector of corresponding individual damping quantities, receiving the frequency vector by an interference variable monitor of the adaptive feed forward control, generating with the interference variable monitor, based on the frequency vector A(ω, c) received from the frequency generator, an expected oscillations vector in the form of x(k+1), where x(k+1)=A(ω, c)x(k)+Le(k), x(k) represents a previously determined expected oscillations vector and Le(k) represents a matrix of learning speeds for adaptation in response to the estimated interference torque, and generating with the interference variable monitor the actuation signal for actuating the electric machine in the form of a desired electric machine output torque, the desired electric machine output torque being in the form of u=Cx(k), where C is a vector related to the frequency vector matrix A(ω, c) by the relationships of $A_i=\left[\begin{array}{cc}0& e^{-{\omega }_i{c}_i{T}_s}\\ -e^{-{\omega }_i{c}_i{T}_s}& 2e^{-{\omega }_i{c}_i{T}_s}\cos \left({\omega }_i\sqrt{1-c_i^2{T}_s}\right)\end{array}\right],C_i=\left[\begin{array}{cc}1& 0\end{array}\right]$ for an ith individual frequency ω.sub.i and ith individual damping c.sub.i in a scanning time T.sub.s between k and k+1 cycles of the adaptive feed forward control, and $A=\left[\begin{array}{ccc}{A}_1& \mathrm{.Math.}& 0\\ \mathrm{.Math.}& \ddots & \mathrm{.Math.}\\ 0& \mathrm{.Math.}& A_N\end{array}\right],C=\left[C_1\mathrm{.Math.}{C}_N\right],$ corresponding to superposition of N sinusoidal oscillations, generating with the electric machine the desired electric machine output torque in response to the actuation signal output by the adaptive feed forward control to the electric machine, and applying the desired electric machine output torque generated by the electric machine to the crankshaft for superimposition on a torque generated by the internal-combustion engine.

4. The method according to claim 3, wherein the frequency vector includes individual frequencies corresponding to expected periodic oscillations associated predefined engine orders.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIGS. 1a and 1b are views of a time-related detail of a rotational speed and a torque, respectively, of an internal-combustion engine of a drive train of a hybrid vehicle for illustrating a rotational imbalance during the operation of the internal-combustion engine;

(2) FIG. 2 is a schematic view of an embodiment according to the invention of an adaptive feed forward control for the reduction of the rotational imbalance of the drive train; and

(3) FIG. 3 is a schematic view of the automatic control concept of an interference variable monitor that can be used according to an embodiment of the invention for the implementation of the adaptive feed forward control.

DETAILED DESCRIPTION OF THE DRAWINGS

(4) The following description is based on a hybrid drive train of a vehicle having an internal-combustion engine and an electric machine. The electric machine is capable of superimposing the torque path of the internal-combustion engine to the driven axle or axles. When an electric machine is mentioned in the present description, the electric machine can optionally be operated by an electric motor or a generator.

(5) As a result of free mass moments of inertia, combustion peaks, and secondary effects of the engine aggregates, the torque supplied by the internal-combustion engine has periodical interference torques. Examples are illustrated in FIGS. 1a and 1b, where a time-related detail of the rotational crankshaft speed n.sub.KW and the torque M.sub.KW existing at the crankshaft are respectively illustrated. The courses of the rotational crankshaft speed n.sub.KW and of the crankshaft torque M.sub.KW illustrated as examples in FIGS. 1a and 1b occur as a result of a constant desired rotational speed. The time-related resolution of slightly less than 0.001 s shows that periodical interferences are superimposed on the rotational speed of the crankshaft as well as the torque M.sub.KW present at the crankshaft. The frequency of these interference signals may either be a multiple of the rotational engine speed n.sub.KW or, independently thereof, assume a constant value. In the following, it is assumed that T.sub.s,i: iε[1, N] is one of N periodical interference signals of the combustor torque with the frequency ω.sub.i and the amplitude A.sub.i:
T.sub.s,i=A.sub.i sin(ω.sub.it+Ø)  (1)

(6) In the following description, only an interference torque present at the crankshaft of the drive train is considered to be an interference signal. The summation of all interference torques Σ.sub.i=1.sup.NT.sub.s,i is called a rotational imbalance. The frequencies of the rotational-speed-dependent interference signals range approximately between
f.sub.s,min≈600 1/min*0.5{circumflex over (=)}5 Hz  (2)
for the 0.5.sup.th engine order when idling and
f.sub.s,max≈8000 1/min*4{circumflex over (=)}533 HZ  (3)
for the fourth engine order at n.sub.KW,max.

(7) Which engine order(s) have to be taken into account for the reduction of the rotational imbalance of the drive train is known and depends on the considered engine. In particular, there is a dependence on the engine order with respect to the number of cylinders, with respect to the arrangement of the cylinders relative to one another (V-arrangement or bank arrangement), as well as on the concrete further development, which influences particularly the amplitude of the interference signals. In the above-indicated formulas (2) and (3), for example, for a given engine arrangement, the 0.5.sup.th and the fourth engine order are considered. In this case, for the 0.5.sup.th engine arrangement, a frequency of f.sub.s,min=5 Hz is obtained for the range inaudible to humans. In contrast, the interference frequency f.sub.s,max=533 Hz determined for the fourth engine arrangement is in the range audible to humans.

(8) The rotational imbalance, i.e. the occurrence of the frequencies computed in Formulas (2) and (3), is to be absorbed as well as possible in the drive train in order to avoid acoustic and component stressing limits.

(9) According to embodiments of the invention, a feed forward control is used for this purpose. The feed forward control uses as a reference, in addition to the rotational speed of the internal-combustion engine, the knowledge of simulating as well as possible frequencies contained in the non-uniformity signals, and extinguishing by way of a suitable actuator. The thus-generated signal is thereby adapted in phasing and amplitude to the interference signal. For this purpose, the feed forward control algorithm is adapted by way of a measured signal that contains the residual interference. This takes place by way of a feed forward control algorithm, which is, for example, based on an interference variable monitor. As an alternative, for example, a so-called adaptive notch filter could also be used. Furthermore, additional implementation possibilities are conceivable.

(10) A schematic representation of the method of operation of a feed forward control is illustrated in FIG. 2. The internal-combustion engine is marked VM and the electric machine is marked EM. A first sensor S1, such as a crankshaft signal generator, and a second sensor S2, such as a rotor position sensor, are shown as further hardware components. In this case, the crankshaft signal generator S1 detects a rotational speed of the internal-combustion engine n.sub.KW. The rotor position sensor S2 detects a rotational speed n.sub.EM of the electric machine.

(11) Based on the measured rotational speed n.sub.KW of the internal-combustion engine VM, a frequency vector ω is formed in the block FG. The block FG represents a frequency generator. The frequency vector ω contains those frequencies that are believed to be in the interference signal. These frequencies may depend on the rotational speed n.sub.KW of the engine VM or may assume independent constant values. The approach for determining the frequency vector will be described in detail in the following.

(12) In block SGB, which represents an interference variable monitor, the system dynamics of the periodical oscillations are described, which are contained in the frequency vector ω, i.e. the respective frequencies of the frequency vector. These are corrected in the SGB with respect to their phase and amplitude by way of a fault signal, are added up, and are provided as a desired torque M.sub.EM,desired as a control signal to the electric machine EM. The electric machine EM will then superimpose the torque M.sub.KW coming from the internal-combustion engine VM with the inverse value of the estimated interference M.sub.EM,actual in order to extinguish the interferences of the rotational imbalance. In the case of an ideal absorption of the rotational imbalance, the resulting output torque M.sub.o and therefore the rotational speed of the electric machine n.sub.EM, which represents the current rotational speed of the electric motor EM, is smooth. A detailed description of the interference variable monitor SGB will follow.

(13) The rotational speed of the electric machine EM measured by the rotor position sensor S2 is fed to an SMS block—an interference torque estimator. From the measured rotational speed n.sub.EM of the electric machine, the interference torque estimator SMS forms an estimation of the interference torque, which is situated on the crankshaft. A conceivable implementation will be further explained below.

(14) When the rotational speed of the electric machine n.sub.EM is smooth under ideal conditions, the interference torque M.sub.e=0. Therefore, no adaptation of the system dynamics takes place in the interference variable monitor SGB. In contrast, should an interference torque M.sub.e be present, the phasing and the amplitude of the oscillations are changed such that the fault is reduced in the next period. A learning speed 1, which is relevant in this case, will be described in detail in the description of the interference variable monitor SGB.

(15) The frequency generator FG and the interference variable monitor SGB form the adaptive feed forward control F. The rotor position sensor S2 and the interference torque estimator SMS represent a measuring path H of a feed forward control. A control loop P, which is not explicitly illustrated in FIG. 2, is provided between the drive torque M.sub.o and the detection of the rotational speed n.sub.EM of the electric machine.

(16) The structure of the interference variable monitor SGB is illustrated in FIG. 3. The equations on which the description of the system dynamics is based are:
x(k+1)=A(ω, c)x(k)+Le(k)  (4)
u=Cx(k)  (5)
wherein x represents a vector with the conditions of the oscillations to be imaged; A represents the matrix for the description of the system dynamics; e represents the remaining residual interference torque; and L represents the matrix of the earning speeds for the adaptation of the conditions to the measured interference dynamics. The vectors ω and c parameterize the system matrix A with respect to the frequency and the damping in the individual sinusoidal oscillations.

(17) The dynamics of an individual vibration with the frequency ω.sub.i and an optional damping c.sub.i in the discrete state space model are described by:

(18) $\begin{array}{cc}{x}_i\left(k+1\right)=A_i{x}_i\left(k\right)& \left(6\right)\\ y_i\left(k\right)=C_i{x}_i\left(k\right)& \left(7\right)\\ \mathrm{with}{A}_i=\left[\begin{array}{cc}0& e^{-{\omega }_i{c}_i{T}_s}\\ -e^{-{\omega }_i{c}_i{T}_s}& 2e^{-{\omega }_i{c}_i{T}_s}\cos \left({\omega }_i\sqrt{1-c_i^2{T}_s}\right)\end{array}\right],C_i=\left[\begin{array}{cc}1& 0\end{array}\right]& \left(8\right)\end{array}$
wherein T.sub.s represents a scanning time. For the superimposition of N sinusoidal oscillations, the following overall matrices are thereby obtained:

(19) $\begin{array}{cc}A=\left[\begin{array}{ccc}{A}_1& \mathrm{.Math.}& 0\\ \mathrm{.Math.}& \ddots & \mathrm{.Math.}\\ 0& \mathrm{.Math.}& A_N\end{array}\right],C=\left[C_1\mathrm{.Math.}{C}_N\right]& \left(9\right)\end{array}$

(20) The advantage of the use of the optional damping c.sub.1 is the reaching of a secure condition, whereby a stable system performance can be ensured. Should such a damping not be provided, c=0 is set, whereby a system of marginal stability is provided.

(21) In FIG. 3, A marks an A-matrix generator to which the frequency ω.sub.i and the damping c.sub.1 are fed. The system matrix for c=0 is illustrated in FIG. 3.

(22) In principle, there is the possibility of adapting conditions to control variable limitations and, as required, switching individual functions to be inactive. For an initial condition x(0), the selection of a vector is recommended that is part of the trajectory of a sinusoidal oscillation, in order to accelerate the transient phenomenon, for example,
x(0)=[1 1 . . . 1].sup.T  (10)

(23) The selection of the learning speeds always has to take place in pairs for the two conditions of a sinusoidal oscillation. In this case, the different frequencies can be weighted differently. Solely, the two conditions of the same frequency have to find the same learning speed:
L=[l.sub.1l.sub.1l.sub.2l.sub.2 . . . l.sub.Nl.sub.N].sup.T  (11)

(24) The selection of l.sub.i has to weigh the stability of the learning algorithm against the convergence time. While the selection of an l.sub.i that is too low results in an insufficient extinction quality, the selection of an l.sub.i that is too large may lead to instability.

(25) Embodiments of the invention therefore provide the use of an adaptive feed forward control in order to absorb or at least reduce the non-uniformity of an internal-combustion engine by way of an electric machine. The adaptive feed forward control utilizes the knowledge that the frequency components to be absorbed depend on the rotational speed of the engine and their frequency is known for any engine order.

(26) In the above description, only an interference variable monitor is described for implementing a solution. It is also easily conceivable to implement such a concept by way of an LMS filter or similar variants.

LIST OF REFERENCE SYMBOLS

(27) VM Internal-combustion engine EM Electric machine S1 Sensor (crankshaft generator) S2 Sensor (rotor position sensor) FG Frequency generator SGM Interference variable monitor n.sub.EM Rotational speed of the electric machine n.sub.KW Rotational speed of the internal-combustion engine M.sub.KW Torque of the internal-combustion engine M.sub.EM.des Desired torque of the electric machine M.sub.EM.act Actual torque of the electric machine M.sub.o Drive torque F Feed forward control H Measuring path

(28) The foregoing disclosure has been set forth merely to illustrate the invention and is not intended to be limiting. Since modifications of the disclosed embodiments incorporating the spirit and substance of the invention may occur to persons skilled in the art, the invention should be construed to include everything within the scope of the appended claims and equivalents thereof.