Battery emulator and method for controlling the battery emulator

Abstract

In order to achieve sufficiently stable output voltage with low losses even during rapid load changes in a battery emulator, a battery emulator is controlled using model-based control with a model of the battery emulator, wherein a line inductance of the electric line and the back-up capacitor is integrated into the model of the battery emulator.

Claims

1. A method for controlling a battery emulator, comprising: a voltage supply having an output filter with a filter capacitor and a back-up capacitor having an output voltage applied, the back-up capacitor connected in parallel to the filter capacitor, wherein the filter capacitor is connected to the back-up capacitor with an electric line in order to physically separate the back-up capacitor from the voltage supply, the method comprising: controlling the output with a model-based control using a model of the battery emulator, wherein a line inductance of the electric line and the back-up capacitor are integrated into the model of the battery emulator.

2. The method according to claim 1, wherein a load model of an electrical test object that is powered by the battery emulator is additionally integrated into the model of the battery emulator.

3. The method according to claim 2, wherein an input capacitance of the test object is used as the load model.

4. The method according to claim 2, wherein a constant power load is used as the load model, wherein the constant power load is linearized around an operating point of the battery emulator.

5. A battery emulator comprising: a voltage supply, having: an input-side rectifier with a direct current intermediate circuit, and a DC-DC converter that is connected to the direct current intermediate circuit, and an output filter with a filter capacitor at the output of the DC-DC converter, a back-up capacitor having an output voltage applied, an electric line connecting the filter capacitor to the back-up capacitor in order to physically separate the back-up capacitor from the voltage supply, and an emulator control unit for model-based control of the output voltage, in which a model of the battery emulator is implemented, wherein a line inductance of the electric line and the back-up capacitor are integrated into the model of the battery emulator.

6. A method of using the battery emulator according to claim 5, comprising testing an electrical test object, wherein the battery emulator is connected to the electrical test object and provides a supply voltage for the electrical test object.

7. The method according to claim 6, wherein a load model of an electrical test object is additionally integrated into the model of the battery emulator.

8. The method according to claim 1, wherein the back-up capacitor is arranged in a junction box physically separate from the voltage supply.

9. The method according to claim 1, further comprising testing an electrical test object, wherein the battery emulator is connected to the electrical test object and provides a supply voltage for the electrical test object.

10. The battery emulator according to claim 5, further comprising a junction box in which the back-up capacitor is arranged, the junction box physically separate from the voltage supply.

11. A battery emulator for testing a test object, comprising: a voltage supply having an output filter with a filter capacitor at the output of a DC-DC converter; a back-up capacitor having an output voltage applied; an electric line connecting the filter capacitor and the back-up capacitor in order to physically separate the back-up capacitor from the voltage supply; an emulator control unit for model-based control of the output voltage, the emulator control unit implementing a model of the battery emulator that integrates a line inductance of the electric line and the back-up capacitor.

12. The battery emulator according to claim 11, further comprising a junction box in which the back-up capacitor is arranged, the junction box physically separate from the voltage supply.

13. The battery emulator according to claim 11, wherein the model of the battery emulator comprises a load model of an electrical test object that is powered by the battery emulator.

14. The battery emulator according to claim 13, wherein a constant power load is used as the load model, wherein the constant power load is linearized around an operating point of the battery emulator.

15. A method of using the battery emulator according to claim 11, comprising testing an electrical test object, wherein the battery emulator is connected to the electrical test object and provides a supply voltage for the electrical test object.

16. A test bench, comprising the battery emulator according to claim 11.

17. A method of using the test bench according to claim 16, comprising: connecting a test object on the test bench to the battery emulator; controlling supply voltage to the test object.

18. The method according to claim 17, wherein the test object comprises a hybrid drive train.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The present invention is explained in greater detail below with reference to FIGS. 1 through 3, which show exemplary, schematic, non-restrictive and advantageous embodiments of the invention. The following is shown:

(2) FIG. 1 a battery emulator according to the prior art,

(3) FIG. 2 a battery emulator according to the invention with an electrical test object and

(4) FIG. 3 a block diagram of the model structure of the battery emulator.

DETAILED DESCRIPTION

(5) The battery emulator 1 according to the present teachings comprises an input-side rectifier 2, which is connected via a direct current intermediate circuit 9 to an intermediate circuit voltage V.sub.0 and an intermediate circuit capacitor C.sub.0 with a DC-DC converter 3. The battery emulator 1 is powered by an alternating current network AC. An output filter 6 comprising a filter inductor L.sub.F in series with the output line and a parallel connected filter capacitor C.sub.F is arranged on the output of the DC-DC converter 3. As is known, the DC-DC converter 3 can also have a multiphase configuration, wherein a filter inductor L.sub.F is provided for each phase in this case. The DC-DC converter 3 is implemented, for example, as a synchronous converter having a plurality of half-bridges (one half-bridge per phase) with semiconductor switches. A battery emulator 1 such as this is known from FIG. 2 of WO 2013/174967 A1, for instance.

(6) On the output side, a back-up capacitor C.sub.S is further provided parallel to the output connectors, to which the output voltage u.sub.A is applied, and parallel to the filter capacitor C.sub.F.

(7) Furthermore, an emulator control unit 5 that controls the DC-DC converter 3 and/or the switches of the DC-DC converter 3 is provided in the battery emulator 1 in order generate the desired output voltage u.sub.Asoll, which is requested by a superordinate control unit. Usually, a pulse width modification PWM is implemented in order to drive the semiconductor switches of the DC-DC converter 3, as is sufficiently well known and as is indicated in FIG. 1. The pulse width modification PWM could also be implemented directly in the emulator control unit 5. Said emulator control unit 5 generates the control variable s for the DC-DC converter 3 or the pulse width modification (PWM) from the set-point value of the control, in this case the output voltage u.sub.Asoll.

(8) As the first measure according to the invention, the back-up capacitor C.sub.S of the battery emulator 1 is physically separated and by distance from the rest of the components of the battery emulator 1. The back-up capacitor C.sub.S is arranged in a separate junction box 7, for example, as is shown in FIG. 2. The result is a distributed battery emulator 1 with a voltage supply 8 and a physically separate junction box 7 with the back-up capacitor C.sub.S. The junction box 7 with the back-up capacitor C.sub.S is then connected to the output filter 7 via a line 4. The rectifier 2, the DC intermediate circuit, the DC-DC converter 3 and the output filter 6 are arranged in the voltage supply 8. This makes it possible, despite the structural size of the battery emulator 1, to arrange the back-up capacitor C.sub.S physically separated from the power supply 8 and close to the electrical load. The line 4 can thus be very long, as is indicated by the interruption in FIG. 2, and can even reach lengths of 10 to 50 m. For this reason, although it is possible to stabilize the output voltage u.sub.A of the battery emulator 1 that is applied to the back-up capacitor C.sub.S, the dynamics of the distributed battery emulator 1 become more complex and difficult to control because the resulting line inductance L.sub.L together with the filter capacitor C.sub.F and the back-up capacitor C.sub.S form an additional oscillating circuit. As a result, this oscillating circuit has another resonance in addition to the one between the filter inductor L.sub.F and the filter capacitor C.sub.F. The controller of the battery emulator 1 may not excite the resonances and it must dampen them when excited by the test object 10.

(9) In order to make it possible to control the battery emulator 1 that is distributed in this way so as to achieve high dynamics (high rate of change in the output voltage u.sub.A), a model-based control based on a model of the battery emulator 1 is provided. The model of the battery emulator 1 is used in the emulator control unit 5 in a model-based control, such as model-predictive control, to control the battery emulator 1. Model-based control in this instance means that the model and/or the model output is used to calculate the control variable s.sub.k of the battery emulator 1 for the next sampling step k. The model of the battery emulator 1 also includes the back-up capacitor C.sub.S and the line inductance L.sub.L of the line 4 that is present between the voltage supply 8 and the back-up capacitor C.sub.S, as is indicated in FIG. 2.

(10) In the application according to the invention, the line inductance L.sub.L is dominant and sufficient. It should be noted, though, that the capacitance per unit length and/or the conductance per unit length and/or the resistance per unit length of the line 4 could additionally be taken into account in the model of the battery emulator 1.

(11) To perform a test run on a test bench 20, the battery emulator 1 and/or the junction box 7 of the battery emulator 1 is connected to the electrical test object 10. Said test object 10 consists e.g. of a drive inverter 11 that powers an electric motor M. The electric motor M propels any load 12, such as a load machine or a drive train with a load machine. The test object 10 can be a hybrid drive train of a vehicle, for instance. A test bench computer 30 is also provided on the test bench 20 to control and monitor the execution of the test run. In so doing, the test bench computer 30 sets the desired output voltage u.sub.Asoll and a set-point value for the drive inverter 11. Measuring devices are also provided on the test bench 20, of course, in order to acquire the required measurement values for control to carry out the test run, such as a torque, a speed, electrical currents or electrical voltages in the hybrid drive train. For the sake of clarity, the measuring devices are not shown.

(12) This results in the block diagram of the model of the battery emulator 1, as shown in FIG. 3, wherein the electrical test object 10 with a load model is also included in this case.

(13) From an electrical standpoint, the test object 10 forms a constant power load CPL, as is described in WO 2013/174967 A1. The constant power load SPL results in a non-linear equation of state, which is linearized around an operating point, as is likewise described in WO 2013/174967 A1. The relationship between the current accommodated by the constant power load CPL and the supply voltage u.sub.A of the constant power load CPL is then expressed as

(14) ${\tilde{i}}_L=\frac{P}{u_A},$
with the power requirement P of the test object 10. By introducing an operating point-dependent differential equivalent resistance

(15) $r_P=-\frac{u_A^2}{P},$
the equation of state can be linearized around an operating point in the form of an output voltage u.sub.A and a load current i.sub.L. This load model in the form of the constant power load CPL can likewise be integrated into the model of the battery emulator 1, as is described in WO 2013/174967 A1. In a simpler configuration, the load model could simply be formed from the input capacitance of the test object 10. This input capacitance can be measured easily or is known. However, the load model also cannot be integrated into the model of the battery emulator 1 at all.

(16) With the model structure as shown in FIG. 3, the following equation of state can be established as a model of the battery emulator 1, in which the input capacitance C.sub.P of the test object 10 is used as a load model, although it could also be omitted for the sake of simplicity. With the state vector x.sub.c=[i.sub.1 v.sub.1 i.sub.2 u.sub.A-].sup.T, which is measured during operating, the state space model is expressed as

(17) $\stackrel{.}{x_c}=\left[\begin{array}{cccc}\frac{-R_{L_F}}{L_F}& \frac{-1}{L_F}& 0& 0\\ \frac{1}{C_F}& 0& \frac{-1}{C_F}& 0\\ 0& \frac{1}{L_L}& \frac{-R_{L_L}}{L_L}& \frac{-1}{L_L}\\ 0& 0& \frac{1}{C_S+C_P}& 0\end{array}\right]x_c+\left[\begin{array}{c}\frac{1}{L_F}\\ 0\\ 0\\ 0\end{array}\right]s+\left[\begin{array}{c}0\\ 0\\ 0\\ \frac{-1}{C_S+C_P}\end{array}\right]i_L.$

(18) Here, R.sub.L.sub.F represents the parasitic resistance of the filter inductor L.sub.F, and R.sub.L.sub.L represents the line resistance of the line 4, which is known or can be measured. The control variable s is obtained from s=d.Math.u.sub.0, with the duty cycle d of the pulse width modulation PWM. In the case of a multi-phase DC-DC converter 3, the individual filter inductors of each phase are combined into a filter inductor L.sub.F, and the currents of the individual phases are added to a common choke current i.sub.1. Naturally, values of the state vector x.sub.c could also be estimated by a control observer if they are not measured directly.

(19) Using the described load model for a constant power load CPL, it is possible to expand this state space model by inserting the differential equivalent resistance r.sub.P.

(20) $\stackrel{.}{x_c}=\left[\begin{array}{cccc}\frac{-R_{L_F}}{L_F}& \frac{-1}{L_F}& 0& 0\\ \frac{1}{C_F}& 0& \frac{-1}{C_F}& 0\\ 0& \frac{1}{L_L}& \frac{-R_{L_L}}{L_L}& \frac{-1}{L_L}\\ 0& 0& \frac{1}{C_S+C_P}& \frac{-1}{\left(C_S+C_P\right)r_P}\end{array}\right]x_c+\left[\begin{array}{c}\frac{1}{L_F}\\ 0\\ 0\\ 0\end{array}\right]s+\left[\begin{array}{c}0\\ 0\\ 0\\ \frac{-1}{C_S+C_P}\end{array}\right]{\tilde{i}}_L$

(21) This equation of state applies to a particular operating point of the battery emulator 1. For this reason, the model must be adapted to the respective operating point during operation. The advantage of utilizing this load model is that only two additional parameters are required for it, which are simple to determine.

(22) For the control, the time-continuous state space model is converted into a time-discrete state space model in a known way. The sampling A is indicated in FIG. 3 and can be carried out e.g. at a frequency of 16 kHz.

(23) The model of the battery emulator 1 with the model parameters can be established beforehand and can be considered known. The load model, on the other hand, can change depending on the electrical load connected to it and is often unknown. In this instance, the model parameters of the load model can be identified by automated identification methods, which are known per se.

(24) To this end, the test set-up, consisting of the battery emulator 1 and the test object 10, can be excited with an excitation sequence in the form of a prescribed time characteristic of the output voltage u.sub.A. If the input capacitance C.sub.P of the test object 10 is less than the capacitance of the back-up capacitor C.sub.S, the identification can be carried out with the disconnected test object 10. If the input capacitance C.sub.P of the test object 10 is greater than or equal to the capacitance of the decoupling capacitor C.sub.S, the test object 10 significantly influences the dynamics and must be connected for parameter identification. However, the parameter identification can then be carried out without a load and with a shut-off test object 10. The reaction of the test set-up is measured and recorded in the form of measured values (according to the model structure). The model of the battery emulator 1 (output filter 8+line 4+junction box 7 with back-up capacitor C.sub.S+test object 10, if required) is subsequently excited with the same excitation sequence, and the model output is simulated and likewise recorded. The difference between the measured values/signals that are measured and the measured values/signals that are simulated is then used as an error in order to minimize these errors in an optimization, e.g. with a least square method, as a function of the model parameters. This identification of the load model can be performed before each test run, for example, or else one time for each test object 10.