Method for damping electromechanical oscillations in an electromechanical system and oscillation damping system for employing such method

Abstract

A method is provided for damping electromechanical oscillations in an electromechanical system including at least one electric machine operable with an angular speed and a phase of the angular speed being coupled to at least one torque load and producing an electromotive force. The method includes deriving actual values of the electromotive force and adjusting the phase of the angular speed, based on the derived actual values of the electromotive force, in a direction in which an oscillatory behavior of the electric machine is reduced.

Claims

1. A method for damping electromechanical oscillations in an electromechanical system comprising at least one electric machine being coupled to at least one torque load and producing an electromotive force, comprising the steps of: deriving actual values of the electromotive force to provide an electromotive force signal; filtering the electromotive force signal in a filter to provide a feed-forward electromotive force signal, the filter providing a phase shift in a frequency region corresponding to a resonance frequency of the electromechanical system; and using the feed-forward electromotive force signal in an open-loop control of the electric machine to damp electromechanical oscillations in the electromechanical system, wherein the step of deriving the actual values of the electromotive force comprises at least one of (i) deriving the values from a calculated magnetic flux of the electric machine, (ii) deriving the values from a calculated electrical angular speed of the electric machine, and (iii) deriving the values from an estimation made by an observer of the state of the electric machine, and wherein the electrical angular speed is calculated from values of a mechanical angular speed of the electric machine and a pole pair number of the electric machine.

2. The method according to claim 1, wherein the step of deriving actual values of the electromotive force is conducted in parallel to and independent from at least one of the further steps (i) of controlling a torque of the electric machine and (ii) of controlling any torque request.

3. The method according to claim 1, (i) wherein parameters of the filter are set according to a simulation of the behavior of the electric machine and/or the electromechanical system or (ii) wherein parameters of the filter are varied adaptively during operation of the electric machine or (iii) where one or more parameters of the filter belonging to a first group of parameters are set according to a simulation of the behavior of the electric machine and where one or more of parameters of the filter belonging to a second group are varied adaptively during operation of the electric machine.

4. The method according to claim 1, wherein the step of using the feed-forward electromotive force signal in an open-loop control of the electric machine comprises providing the feed forward signal as input to a current control segment in addition to at least one of a requested current derived from a corresponding torque request an electrical angular speed of the at least one electric machine, an actual current of the at least one electric machine.

5. The method according to claim 1, further comprising the step of forwarding control signals to an inverter coupled to the at least one electric machine.

6. An oscillation damping torque control system for an electric machine coupled to at least one torque load and producing an electromotive force, comprising a current controller connected to received feedback from the electric machine and configured to apply current control of the electric machine, and a filter connected to receive an electromotive force signal derived from actual values of the electromotive force and to provide a feed forward electromotive force signal to the controller, the filter providing a phase shift of the electromotive force signal in a frequency region corresponding to a resonance frequency of the electromechanical system, wherein the current controller is further configured to apply the feed forward electromotive force signal in an open loop control of the electric machine to damp electromechanical oscillations in the electric machine, wherein the actual values of the electromotive force are derived from at least one of: (i) a calculated magnetic flux of the electric machine, (ii) a calculated electrical angular speed of the electric machine, and (iii) an estimation made by an observer of the state of the electric machine, and wherein the electrical angular speed is calculated from values of a mechanical angular speed of the electric machine and a pole pair number of the electric machine.

7. The system according to claim 6, further comprising a torque control segment configured to receive a requested torque and to transform it into a requested current, and to provide the requested current to the current controller.

8. The system according to claim 6, wherein the electric machine is coupled to one or more torque loads characterized by a mechanical angular speed via a spring-damper system.

9. A computer comprising a computer program stored on a non-transitory computer readable medium adapted to perform a method or for use in a method according to claim 1.

10. The computer according to claim 9 connected to the internet and arranged to download the program to a control unit or one of its components.

11. Computer program product stored on a non-transitory computer readable medium, comprising a program code for use in a method according to claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The present invention together with the above-mentioned and other objects and advantages may best be understood from the following detailed description of the embodiments, but not restricted to the embodiments, wherein is shown schematically:

(2) FIG. 1 an example of an electromechanical drive system according to the invention;

(3) FIG. 2 angular speed curves with and without filtering of the angular speed;

(4) FIGS. 3a, 3b an example of amplitude and phase of a phase advancing filter represented by a Bode diagram (FIG. 3a) and two signals as a function of time indicating a phase advancement between an input signal to and an output signal of a phase advance filter (FIG. 3b);

(5) FIG. 4 a flow chart illustrating a filter parameter estimation problem;

(6) FIG. 5 a first embodiment for applying the method according to the invention employing pre-defined filter parameter setting;

(7) FIG. 6 a second embodiment for applying the method according to the invention employing a filter parameter setting based on gain scheduling;

(8) FIG. 7 a third embodiment for applying the method according to the invention employing an adaptive filter parameter setting; and

(9) FIG. 8 a sketch of a drive system of a hybrid vehicle incorporating a system according to the invention.

DETAILED DESCRIPTION

(10) In the drawings, equal or similar elements are referred to by equal reference numerals. The drawings are merely schematic representations, not intended to portray specific parameters of the invention. Moreover, the drawings are intended to depict only typical embodiments of the invention and therefore should not be considered as limiting the scope of the invention.

(11) FIG. 1 depicts schematically an example embodiment of an electromechanical system 50 comprising a controllable electric machine 20. The electromechanical system 50 may be an electromechanical drive train of an automotive hybrid vehicle (schematically shown in FIG. 8). The electromechanical system 50 comprises a DC-voltage circuit 10 with a battery providing a battery voltage U_bat and supplying a supply voltage U_dc to an inverter 12. The inverter 12 supplies a current I to the electric machine 20. The DC-voltage circuit 10 supplies a DC voltage U dc to the electric machine 20. Control signal/sensor signal lines (indicated in form of arrows) are connecting a control unit 100 with the DC-voltage circuit 10, the inverter 12 and the electric machine 20. The control unit 100 controlling the DC-voltage circuit 10 senses the DC voltage U dc for controlling purposes. The DC-voltage supplied to boxes 102 and 104 indicates the measured voltage used for internal voltage U_dc used for internal limitations, such as saturation functionality, calculation of field weakening in the torque control, which is well known in the art.

(12) In the Figure, a mechanical load connected to the electric machine 20 is modelled as a spring-damper unit 30 and a torque load 40 (for instance a combustion engine) as a general representation of any mechanical load. Other representations may be possible depending on the model used for describing such an electromechanical system. The electric machine 20 is mechanically coupled to the torque load 40 (such as the combustion engine (not shown)) by the spring-damper unit 30 which represents the mechanical and torsional properties as well as eventual damping properties of the connection between the electric machine 20 and the torque load 40.

(13) The electric machine 20 comprises an electrical part 22 represented by an inductance L and a resistance r of the electric machine 20 and a mechanical part 24 represented by an inertia system of the rotating parts of the electric machine 20.

(14) The inverter 12 receives input signals from the control unit 100. The electric machine 20 may be for instance a synchronous machine, e.g. a Permanent Magnet (PM) electric machine.

(15) The control unit 100 encompasses several segments which can be embodied as hardware and/or as software in the control unit 100. Particularly, the control unit 100 may comprise a torque control segment 102, a controller 104 which contains a current control segment and a modulation segment (not shown in details), a filter 110 and a pole pair number segment 112 which contains information on the pole pair number of the electric machine 20.

(16) The output of the control unit 100 is a modulated voltage signal depending on the inverter design, for instance a pulse width modulated signal (PWM signal), a pulse frequency modulated signal (PFM signal) or a pulse step modulated signal (PSM signal).

(17) The torque control segment 102 receives a torque request M_req, e.g. derived from an accelerator position, and outputs a current signal I req to the controller 104 of the control unit 100. The accelerator may be an accelerator or gas pedal of the hybrid vehicle.

(18) Additionally, the torque control segment 102 receives the mechanical angular speed jmech and the electrical angular speed _el as input signals. The electrical angular speed _el is a result of the mechanical angular speed _mech combined with the pole number of the electric machine 20 provided by pole pair number segment 112.

(19) The torque control segment 102 may include a unit for achieving a field weakening of the electric machine 20 and includes limitations with respect to current and voltage applicable to the electric machine 20. Thus, the requested torque M_req is transformed by the torque control segment 102 to the requested current I_req inputted to the controller 104. By use of such control object as field weakening a reversed magnetic flux can be introduced which weakens a magnetic flux from the permanent magnets at high speed, which can be achieved by requesting a reactive electric current in the windings of the electric machine 20. If field weakening is not performed the electromotive force generated by the electric machine 20 may yield a voltage for the inverter 12 which might under certain conditions be too high for the converter 12 to control.

(20) Additional to the current input I_req, the controller 104 receives as input a measured current I_meas of the electric machine 20 and the electrical angular speed el and a feed forward signal of the electromotive force emf_ff of the electric machine 20 supplied by the filter 110. The electromotive force signal emf_ff is used as a feed forward signal in the sense that the phase is adjusted in an open loop control instead of a closed loop control.

(21) It should be noted that the controller 104 may act as a feedback controller for one or more input parameters such as the measured current I_meas and the like but that the electromotive force signal emf_ff is used as a feed forward signal inputted to the controller 104 so that the phase of the signal is not subject to feedback but by feed forward only.

(22) The electromotive force signal emf_ff is derived from the mechanical angular speed mech and the pole pair number of the electric machine 20 and can either be calculated or can be estimated by use of a state observer. Input parameters for the electromotive force signal emf_ff are expediently the mechanical angular speed mech, the pole pair number and the permanent magnetic flux _m. It is possible, however, to use other parameters for estimation of the electromotive force signal emf_ff, depending on the model used, such as e.g. a change in the mechanical angle of the current in the electric machine 20, the geometry of the electric machine 20 and the rotor flux to make a similar estimate.

(23) A state observer is a system that models a real system in order to provide an estimate of its internal state, given measurements of the input and output of the real system. It is typically a computer-implemented mathematical model. In case the electric machine 20 is a PM electric machine, the electromotive force emf_ff is for instance simply the product of the pole number and the mechanical angular speed jnech, the permanent magnetic flux _m and a filter function F.

(24) The property of the phase shift is determined by the particular resonance frequency to be damped and the associated eigenvalue.

(25) The filter 110 causes generally a phase shift of a signal which is fed into the filter 110. Particularly, the phase in a plot of a frequency versus a phase of the system can be shifted if for instance an oscillation of the mechanical angular speed _mech is observed in a model simulation of the electromechanical system.

(26) Filter 110 can be a filter with fixed parameters which were set in advance according to the resonance behaviour of the electric machine 20. Filter 110 can be considered as a feed forward filter. In this case the phase of the signal is shifted (by way of example) by a constant phase when processed by the filter 110. The filter parameters can be generated by model calculations in a design phase of the filter 110 and/or the control unit 100. Alternatively, the filter 110 can be an adaptive filter which changes its parameters during operation according to operation conditions of the control system 100 and/or the electric machine 20 and the interconnected mechanical system.

(27) FIG. 2 depicts angular speed curves as a function of time and illustrates the differences in mitigation of oscillations. Electromechanical systems are frequently prone to exhibit torsional resonances that may cause torsional oscillations. These oscillations are mainly caused by mechanical spring-dampers that should mitigate higher frequency torque alternations to propagate through-out the system. However, in conjunction with an electrical machine 20 that is controlled by an electrical control device 100 connected to an electrical system, the mechanical system may interfere with the electrical system and the controls thereof. The intrinsic mechanical system damping properties may be abated as a consequence. This effect is demonstrated by curve A in FIG. 2 showing a rather strong oscillating behaviour of the mechanical angular speed _mech of the electric machine 20. Particularly, curve A represents a case where no filter 110 (FIG. 1) is applied. Instead, some minor inherent damping is present resulting in a slight reduction of the oscillation amplitude with increasing time.

(28) If the mechanical angular speed _mech of the electric machine 20 is strongly oscillating (as in curve A in FIG. 2), the electromotive force emf ff, being proportional to the product of mechanical angular speed _mech, pole pair number and permanent magnetic flux _m, will exhibit such oscillations, too.

(29) However, according to the invention, by adjusting the phase of the mechanical angular speed jnech by use of the filter 110 (FIG. 1) it is possible to abate such oscillations. This effect is demonstrated by curve B in FIG. 2 showing, after a short start phase, a rather smooth curve whereby even slight oscillations in the mechanical angular speed jnech are virtually eliminated shortly after the start phase. As curve B in FIG. 2 shows, the unwanted oscillations decrease very quickly approaching virtually to zero after a rather short period of time.

(30) For illustrating of the invention, FIG. 3a and FIG. 3b show the effect of filter 110 (as shown in FIG. 1). FIG. 3a depicts a so called Bode diagram. Generally a Bode diagram consists of two graphs identifying a complex transfer function, a graph indicating an absolute value of an amplification of an amplitude (called magnitude in the upper diagram in FIG. 3a) on a logarithmic scale, and a graph for an argument (called phase in the lower diagram in FIG. 3a) of the complex transfer function, both graphs plotted versus a frequency, plotted with logarithmic frequency axes, to show the transfer function or frequency response of a linear, time-invariant system, thus showing a stationary response of an output of a system to a harmonic input to the system.

(31) In FIG. 3a the magnitude is displayed in units of dB, the phase is displayed in units of degree and the frequency is displayed in units of rad/sec. The Bode diagram in FIG. 3a shows that for frequencies above a certain frequency value these frequencies are magnified (i.e. the signals are increased in amplitude) but also that these frequencies are phase advanced. For a filter (110 in FIG. 1) the Bode diagram illustrates the filter parameter setting in the frequency plane. For the filter depicted in FIG. 3a, the magnitude (amplitude) of the filtered signal increases as a function of frequency while the phase of the filtered signal increases as a function of frequency until a maximum and then, at higher frequencies, decreases again.

(32) Hence, as shown in FIG. 3b, an input signal that exhibits some frequency component which could be caused by a resonance, within this typical frequency region of the filter (110 in FIG. 1), will be phase advanced.

(33) FIG. 3b shows by way of example the amplitude A of two sinusoidal signals S1 and S2 as a function of time t. The Figure shows how a signal S1 of a certain frequency will be affected by a filter with phase advancing characteristics resulting in the phase advanced signal S2. However, when such filter is applied within an electromechanical system 50 comprising a controllable electric machine 20 according to FIG. 1, the impact on the system 50 (FIG. 1) will be an attenuation effect (as shown by curve B in FIG. 2) that will be superimposed to the phase advancing effect of the filter.

(34) More particularly, the amplitudes AMP of the two sinusoidal signals S1, S2 in FIG. 3b as a function of time t indicate a phase advancement between the input signal S1 to the filter with a phase advancing characteristic and the output signal S2 of said filter with filter characteristics as described by the Bode diagram in FIG. 3a. The phase advancement is seen as the corresponding time deviation between the input signal S1 and the filtered output signal S2 of the input signal S1, which corresponds to a phase displacement of the input signal S1, indicated by an arrow in FIG. 3b.

(35) FIG. 4 shows a flow chart illustrating a filter parameter estimation which is a possible basis for the embodiment of the invention shown in FIG. 1. The filter parameter estimation is done in a model phase in which filter parameters are calibrated before the implementation of the filter 110 in the real electromechanical system.

(36) In step 200 in a simulation the system describing the electric machine 20 (FIG. 1) is excited from external disturbances or by a signal injection. The external disturbances can be introduced e.g. from the mechanical system connected to the electric machine 20 (FIG. 1).

(37) In step 202 it is checked if any non-damped or poorly damped resonances are to be observed in the simulated system. If the answer is no (n in step 202) the normal operation is continued in step 204 until any disturbance, any unwanted excitation of the system or any change of work point of operation is observed.

(38) Then the procedure jumps directly back to step 200. If the answer in step 202 is yes (y in step 202), the procedure continues with step 206, where, based on a concurrent analysis of the effect of the excitation on the system resonances, new filter parameters are estimated or calculated to achieve damping properties of the system. After this step, the routine continues with step 200.

(39) FIGS. 5 to 7 illustrate three different embodiments how the invention can be employed, namely (i) with predefined filter parameters, (ii) with gain scheduling and (iii) with an adaptive filter.

(40) Particularly, FIG. 5 depicts a block diagram of the action of a filter 110 according to a first embodiment of the invention, employing pre-defined filter parameter setting up the actual filter 110.

(41) With reference to FIG. 1 and the components and their reference numerals described therein, the mechanical angular speed mech of the electric machine 20 is inputted into a disturbance model 300 which simulates the electromotive force emf ff created by the electric machine 20. The mechanical angular speed jmech can be measured or estimated, the latter possibility allowing the number of necessary sensors to be reduced. The disturbance model 300 comprises the pole pair number 112 and the filter 110 characterized by a filter function F and the permanent magnetic flux _m of the electric machine 20.

(42) The result of the disturbance model 300 is an input for a system 50 comprising an electrical machine 20 and mechanical parts 30 and 40 (as depicted in FIG. 1), wherein the input parameter for the system is a reference voltage U_ref. A controller 104 feeds the output signal of the system parts 20, 30, 40, i.e. a measured current Ijmeas of the system parts 20, 30, 40 back to the input of said system parts 20, 30, 40.

(43) The filter 110 shifts the phase of the input signal jnech according to the predetermined fixed filter parameter set determined during a model phase of the electromechanical system as indicated in FIGS. 2 and 3.

(44) In this embodiment, the filter 110 is designed to mitigate a certain resonance of the electric machine 20 known in advance in the development process designing the electromechanical system.

(45) FIG. 6 depicts a block diagram of a second embodiment for applying the method according to the invention, employing a filter parameter setting based on gain scheduling. In the method of gain scheduling different sets of control parameters can be chosen dependent on operation points. For instance, in case of a vehicle for each gear shift and/or gear applied, during operation of the vehicle a parameter set can be read from a look up table which can be produced during an instant simulation of the electromechanical system.

(46) With reference to FIG. 1 and the components and their reference numerals described therein, the mechanical angular speed jnech of the electric machine 20 is inputted into a disturbance model 300 which simulates the electromotive force emf_ff created by the electric machine 20. The mechanical angular speed _mech can be measured or estimated, the latter possibility allowing the number of necessary sensors to be reduced. The disturbance model 300 comprises the pole pair number 112 and a filter 110 characterized by a filter function F and the permanent magnetic flux _m of the electric machine 20.

(47) The result of the disturbance model 300 is an input for a system 50 comprising an electrical machine 20 and mechanical parts 30, 40 (as depicted in FIG. 1), wherein the input parameter for the system is a reference voltage U_ref. A controller 104 feeds the output signal of the system 20, 30, 40, i.e. a measured current ljneas of the system 20, 30, 40 back to the input of system 20, 30, 40.

(48) The signal experiences a variable phase shift in the filter 110. The phase shift is varied depending on a predefined set of parameters for different system work points of operation provided by a gain scheduling block 308. The gain scheduling block 308 receives a work point of operation W op of the electric machine 20 as input, e.g. from a lookup table which contains parameters for operation modes known in advance. For instance, for each gear shift an appropriate parameter set is read and fed into the filter 110, thus altering the filter function F according to the actual operation mode of the electromechanical system (50 in FIG. 1).

(49) If different operational modes alter the resonance frequencies of the electric machine 20, different sets of control parameters can be used to impose the mitigation. Which set of control parameters should be chosen depends on the actual mode of operation that in this case is known in advance and determined in the development process designing the electromechanical system.

(50) FIG. 7 depicts a block diagram of a third embodiment for applying the method according to the invention, employing an adaptive filter parameter setting.

(51) With reference to FIG. 1 and the components and their reference numerals described therein, the mechanical angular speed _mech of the electric machine 20 is inputted into a disturbance model 300 which simulates the electromotive force emf ff created by the electric machine 20. The mechanical angular speed mech can be measured or estimated, the latter possibility allowing the number of necessary sensors to be reduced. The disturbance model 300 comprises the pole pair number 112 and a filter 110 characterized by a filter function F and the permanent magnetic flux _m of the electric machine 20.

(52) The result of the disturbance model 300 is an input for a system comprising an electrical machine 20 and mechanical parts 30, 40 (as depicted in FIG. 1), wherein the input parameter for the system is a reference voltage U_ref. A controller 104 feeds the output signal of the system parts 20, 30, 40, i.e. a measured current I_meas of the system parts 20, 30, 40 back to the input of the system 20, 30, 40.

(53) A filter parameter estimation process block 310 receives input from the output of the system parts 20, 30, 40. The output of the filter parameter estimation process block 310 is combined with the combined output signals of the controller 104 and the filter 110 forming the input of the system parts 20, 30, 40. The filter parameter estimation process block 310 calculates a loss function for certain operation modes and calculates parameters to minimize these losses. In this embodiment, input and output parameters in the electromechanical system (50 in FIG. 1) are measured and/or estimated, and the filter function F of the filter 110 is adaptively adjusted, wherein an optimization algorithm calculates online the control parameters to be used for resonance mitigation, more particularly the filter function F of the filter part of filter 110.

(54) A schematic representation of a vehicle 90 which employs the method according to the present invention is shown in FIG. 8. The vehicle 90 comprises in a well-known manner wheels 140 and 142 in the back and the front of the vehicle 90. A drive axle 132 is coupled to an internal combustion engine 130 via an electric machine 20. The drive axle 132 is coupled to the wheels 140. The vehicle wheels 140, 142 and the engine 130 are main parts in the mechanical system of the vehicle drive train (represented by numeral 40 in FIG. 1). The electrical machine 20 is controlled by a control unit 100 which employs the method according to the invention.