DE102020104314A1 - Test bench for testing a drive component with a shaft using a model to calculate a future value of a state variable of the shaft - Google Patents

Abstract

The invention relates to a test bench (1) for testing a drive component (2) for a vehicle (3), the test bench (1) has a shaft (4), an electric motor (5) which is connected to the drive component (2) via the shaft (4) can be coupled mechanically, and a control device (6) with a kinematic model (7) for modeling a kinematic system (8) which has at least the drive component (2), the electric motor (5) and the shaft (4), The control device (6) is set up and designed to use the kinematic model (7) to calculate at least one future value of a state variable of the shaft (4) for a future period of time and, depending on the future value of the state variable of the shaft (4), the state variable to regulate the shaft (4).

Description

The invention relates to a test bench for testing a drive component for a vehicle having a shaft and an electric motor which can be mechanically coupled to the drive component via the shaft. The invention further relates to a drive system for a vehicle and a method for testing a drive component for a vehicle.

The WO 2011038429 A1 describes a test bench for emulating the driving behavior of a vehicle with a drive train with at least one shaft on which a loading machine can be mounted. A torque measuring device for measuring the torque on the shaft is connected to a computing device having a wheel model, which is set up to deliver a speed setpoint value for a speed control assigned to the loading machine from the measured torque.

A test bench for testing a drive component for a vehicle is proposed. The test bench has a shaft, an electric motor that can be mechanically coupled to the drive component via the shaft, and a control unit with a kinematic model for modeling a kinematic system. The kinematic system has at least the drive component, the electric motor and the shaft. The control device is set up and designed to use the kinematic model to calculate at least one future value of a state variable of the shaft for a future period and to regulate the state variable of the shaft as a function of the future value of the state variable of the shaft. In the case of such a control, the state variable of the shaft is preferably a controlled variable.

In order to regulate the state variable of the shaft, the control unit advantageously has a controller which, depending on a reference variable, which can be implemented in the form of a setpoint of the state variable of the shaft, and a value of the state variable of the shaft, a value of a manipulated variable for adjusting the State variable of the wave calculated. The control device is preferably set up to send the value of the manipulated variable to the drive component or the electric motor. The drive component, the shaft and the electric motor preferably form a controlled system if the drive component is kinematically coupled to the electric motor via the shaft. The manipulated variable acts on the controlled system for adjusting the value of the state variable of the shaft. The shaft can be designed as an output shaft of the electric motor.

Based on the future value of the state variable calculated using the kinematic model, the behavior of the test bench in the future period can be estimated with the control unit. Because the control device is set up and designed to regulate the state variable as a function of the future value of the state variable, the control unit provides model-predictive control. The model predictive control can in particular increase the control quality of a control of the state variable of the shaft with the aid of the control device.

The control unit is advantageously set up and designed to calculate a future value of the manipulated variable for regulating the state variable of the shaft for the future period, preferably with the aid of the controller. The control unit preferably determines the future value of the manipulated variable using an optimization process. The optimization method advantageously includes a minimization of a cost function. The cost function preferably weights at least the future value of the state variable of the shaft, preferably several future values of state variables of the kinematic system, compared to the future value of the manipulated variable, preferably compared to several future values of the manipulated variable.

The test stand advantageously has fastening elements for fastening the drive component to the test stand, such as threads for screwing on the drive component. The test stand can assume a test state in which the drive component is attached to the test stand and is mechanically coupled to the electric motor via the shaft. By mechanical coupling is meant a coupling in which a torque acting on the shaft from the drive component causes a torque acting on the electric motor from the shaft and a torque acting on the shaft by the electric motor causes a torque generated by the shaft acts on the drive component. The kinematic model is set up to simulate the kinematic system in a state in which the drive component is mechanically coupled to the electric motor. If the drive component is not attached to the test bench, the test bench is in an equipable state.

The vehicle can be in the form of a passenger car, a truck, a rail vehicle, a ship or any vehicle that can be driven using a torque generated by the drive component. The drive component is preferably in the form of a Internal combustion engine trained. The electric motor can be a dynamometer. The state variable of the shaft can be, for example, a speed or a torque acting on the shaft.

If the state variable of the shaft is the speed of the shaft, the control device is preferably set up and designed to regulate the speed of the shaft using the future value of the speed. This can make it easier to operate the drive component at the specified setpoint speeds using the test bench, with a respective current value of the shaft speed being able to have a smaller deviation from the corresponding setpoint speed compared to a control system without calculating the future value of the state variable. As a result, a map of the drive component for a control unit of a vehicle can be created with a comparatively high precision.

The test stand is advantageously designed to apply a torque acting on the shaft from the electric motor, hereinafter referred to as dynamometer torque, using the electric motor. The dynamometer torque is preferably the manipulated variable for regulating the state variable, in particular the speed, of the shaft. The control unit is advantageously set up and designed to control or regulate the electric motor in such a way that the dynamometer torque is applied to the shaft. The testing of the drive component can include operating the drive component at one of the predetermined target speeds.

In a preferred embodiment, the control device is set up and designed to calculate at least one future value of a disturbance variable of the kinematic system for the future period and to regulate the state variable of the shaft as a function of the future value of the disturbance variable. The disturbance variable can act on the controlled system in such a way that it changes the value of the state variable of the shaft.

If the state variable of the shaft is controlled as a function of the future value of the disturbance variable, the control quality of the regulation of the state variable of the shaft can be increased further using the control unit.

In an advantageous embodiment it is provided that the test stand has a sensor for detecting a measured value of a state variable of the kinematic system. As part of this configuration, the control device is set up and designed to estimate a value of the state variable of the kinematic system and to correct the estimated value of the state variable of the kinematic system with the aid of the measured value of the state variable of the kinematic system. Furthermore, in this embodiment, the control device is set up and designed to determine a corrected estimated value of the state variable of the kinematic system as a result of the correction and to calculate the future value of the state variable of the shaft as a function of the corrected estimated value of the state variable of the kinematic system. This makes it possible to take current sensor data, such as the measured value of the state variable of the kinematic system, into account when regulating the state variable of the shaft. This can further increase the control quality.

An advantageous development provides that the control device is set up and designed to take into account an influence of the disturbance variable on the kinematic system when carrying out the correction of the estimated value of the state variable of the kinematic system. If the correction is carried out in this way, the control unit provides a disturbance variable observer with whom the stationary control error can be compensated. This is particularly advantageous in connection with the model predictive control, because the model predictive control can cause stationary control errors to occur. Furthermore, a transient control error can be reduced if the influence of the disturbance variable is taken into account when performing the correction.

The stationary control error can thus be compensated and the transient control error can be reduced by the control device regulating the state variable of the shaft also as a function of the future value of the disturbance variable. Furthermore, the optimization method for determining the future value of the manipulated variable can thereby be improved.

The terms disturbance variable, manipulated variable, controlled variable, command variable, controlled system and controller each denote the term interfering variable, manipulated variable, controlled variable, command variable, controlled system or controller as it is used as standard in the field of control technology and in particular according to DIN 19226 “Control technology, control technology and control technology “Is standardized. In most cases, a value of the disturbance variable cannot be modeled or cannot be calculated for the future period. In particular, the disturbance variable differs from the controlled variable in that the kinematic model considers the controlled variable but not the disturbance variable.

In one development, the control device is set up and designed to calculate the future value of the state variable of the shaft as a function of at least one dead time of the test bench. The dead time can be caused by delays in communication of signals that are sent to the control unit. For example, a speed sensor can scan positions of the shaft and send them to the control unit, the control unit being able to receive them with a signal delay. Furthermore, the dead time can occur when the manipulated variable is set, for example due to required computing times in digital data processing, which can take place, for example, in the control unit or a frequency converter. The control unit advantageously calculates a value of the cost function as a function of the dead time. As a result, the optimization method for determining the future value of the manipulated variable can be further improved.

In an advantageous embodiment, the control device is set up and designed to use the kinematic model to determine the state variable of the shaft as a function of the torque that can be generated on the shaft using the electric motor, i.e. regulating the dynamometer torque and determining the torque that can be generated with the aid of the electric motor as a function of a simulation of at least one component of a drive train of the vehicle. In this embodiment, the control unit can comprise a computing unit for simulating at least the component of the drive train. For example, the control unit can be set up to determine a transmission speed or a time course of the transmission speed using the simulation of the component of the drive train and to calculate the dynamometer torque as a function of the transmission speed or the time course of the transmission speed.

The computing unit is advantageously set up to simulate the component, several components or all components of the drive train, preferably taking into account the respective moments of inertia of the components of the drive train and damping elements of the drive train.

Because the control unit determines the dynamometer torque as a function of the simulation, the drive component can be tested with the test bench in the form of an “engine in the loop” test procedure. This can reduce hardware expenditure to the extent that the component of the drive train does not have to be provided in the form of hardware in order to test or optimize the interaction of the drive component with the component of the drive train. The drive train component may be, for example, a transmission, a flywheel, a clutch, a differential, or a wheel.

The kinematic model can be stored in the control unit in the form of a calculation module. The calculation module is preferably set up to calculate the future value of the state variable of the shaft, preferably as a function of a mass inertia of the drive component and a mass inertia of the electric motor, as a function of a first value of a state variable or first values of state variables of the kinematic system. The first value of the state variable or the first values of the state variables of the kinematic system relate to a first period of time that is temporally ahead of the future period. The future period and / or the first period can be infinitesimally small and be in the form of a future point in time or first point in time.

According to an advantageous embodiment, the control device is set up and designed to determine the future value of the state variable of the shaft as a function of an estimated torque of the drive component and to determine the estimated torque of the drive component using a characteristic diagram of the drive component. As described below, the estimated torque can be determined in the form of a theoretical torque, preferably as a function of an accelerator pedal position, using the characteristic diagram. The map can be determined by previous test bench tests or using a database. The database advantageously includes a set of characteristic maps of further drive components that are mechanically and thermally similar to the drive component.

A calculation of the future value of the state variable of the shaft using the characteristic diagram has the advantage that the torque of the drive component can be determined in the form of the estimated torque without having to solve a differential equation as when carrying out a simulation. This makes it possible to determine the future value of the state variable of the wave in real time. This can further increase the control quality of the control of the state variable of the shaft with the aid of the control device.

The characteristic diagram advantageously maps values of state variables of the drive component to a value of the estimated torque. One of the state variables of the drive component can be a rotational speed of the drive component, in particular a rotational speed of a crankshaft of the drive component. The speed of the drive component is preferably equal to the speed of the shaft if the drive component is mechanically coupled to the electric motor via the shaft.

A drive system for a vehicle is also proposed. The drive system has a drive component, a shaft, an electric motor that is mechanically coupled to the drive component via the shaft, and a control device with a kinematic model for modeling a kinematic system. The kinematic system has at least the drive component, the electric motor and the shaft. The control unit of the drive system is set up and designed to use the kinematic model to calculate at least one future value of a state variable of the shaft for a future period and to regulate the state variable of the shaft as a function of the future value of the state variable of the shaft. Because the control unit of the drive system is set up to take into account the future value of the state variable when regulating the state quantity of the shaft, a behavior of the drive system can be estimated in the future period and thereby a control quality can be increased. The control unit of the drive system can differ from the control unit of the test bench. In a different variant, the control unit of the drive system is the same as the control unit of the test bench.

Furthermore, a method for testing a drive component for a vehicle is proposed. The drive component can be mechanically coupled to an electric motor via a shaft. The procedure has the following steps. In a first step, a future value of a state variable of the shaft for a future period is calculated using a kinematic model for modeling a kinematic system. The kinematic system has at least the drive component, the electric motor and the shaft.

In a second step, the state variable of the shaft is regulated depending on the future value of the state variable of the shaft. Because the state variable of the shaft is regulated as a function of the future value of the state variable of the shaft in the proposed method, the method can achieve the same advantages as with the proposed test bench. The drive components can be checked, for example, in such a way that the rotational speed of the shaft is regulated and, when the rotational speed is reached, state variables of the drive component are measured and compared with limit values. The state variables can be concentrations of pollutants in an exhaust gas of the drive component, for example.

Both the kinematic model of the control unit of the drive system and the kinematic model as used in the proposed method can be designed according to one of the variants of the kinematic model of the control unit of the test bench described above or below.

A computer program product is also proposed. The computer program product comprises a program which, when executed by a computer, causes the computer to carry out a method according to one of the configurations described above or below.

• 1 ein Fahrzeug mit einer Antriebskomponente,
• 2 einen Prüfstand zum Prüfen der in 1 gezeigten Antriebskomponente mit einem Elektromotor, einer Welle und einem Steuergerät mit einem kinematischen Modell,
• 3 eine schematische Skizze einer ersten Ausführungsform des in 2 gezeigten kinematischen Modells,
• 4 einen Regelkreis mit einem Regler zur Regelung einer Zustandsgröße der in 2 gezeigten Welle,
• 5 Schritte eines Verfahrens zum Prüfen der in 1 gezeigten Antriebskomponente,
• 6 ein Antriebssystem für ein weiteres Fahrzeug.

The dependent claims describe further advantageous embodiments of the invention. Preferred exemplary embodiments are explained in more detail with reference to the following figures. It shows schematically

• 1 a vehicle with a drive component,
• 2nd a test bench for testing the in 1 shown drive component with an electric motor, a shaft and a control unit with a kinematic model,
• 3rd is a schematic sketch of a first embodiment of the in 2nd shown kinematic model,
• 4th a control loop with a controller for regulating a state variable in 2nd shown wave,
• 5 Steps of a procedure for checking the in 1 drive component shown,
• 6 a drive system for another vehicle.

1 shows a vehicle 3rd with a drive component 2nd for driving the vehicle 3rd . In a preferred embodiment, the vehicle 3rd designed as a motor vehicle and the drive component as an internal combustion engine.

2nd shows a test bench 1 for testing the drive component 2nd for the vehicle 3rd . The test bench 1 points a wave 4th , an electric motor 5 with the drive component 2nd over the wave 4th is mechanically coupled, and a control unit 6 with a kinematic model 7 for modeling a kinematic system 8th on. The kinematic system 8th has at least the drive component 2nd , the electric motor 5 and the wave 4th on. The control unit 6 is set up and trained using the kinematic model 7 at least one future value of a state quantity of the wave 4th to calculate for a future period and depending on the future value of the state quantity of the wave 4th the state quantity of the wave 4th to regulate. The wave 4th is preferably designed in the form of a cardan shaft and couples the electric motor 5 mechanically via a clutch 10th and a vibration damper 11 with a crankshaft 12th the drive component 2nd .

At the in 2nd The embodiment shown is the electric motor 5 trained in the form of a dynamometer and with the help of a frequency converter 9 controllable. The frequency converter 9 is preferably set up coils of the electric motor 5 to energize such that the electric motor 5 the wave 4th applied with a dynamometer torque. The dynamometer torque can be generated with the aid of respective control currents which flow through the respective coils of the dynamometer. The control currents are provided by the frequency converter 9 advantageous depending on one with the help of the control unit 6 calculated target torque M _{Dyno, s} and one with the help of the control unit 6 calculated target speed n _s .

3rd shows a schematic sketch of a first possible variant of the kinematic model 7 for modeling the kinematic system 8th . The kinematic model 7 in this embodiment is advantageously in the form of a calculation module 30th with individual calculation units 31 , 35 , 38 , 43 , 45 , 47 educated. A first calculation unit preferably calculates 31 depending on an accelerator pedal position 32 of the vehicle 3rd , preferably using a map 33 , a theoretical torque 34 the drive component 2nd . The theoretical torque 34 can be a target torque M _{VM, s of} the drive component 2nd be considered as it is from the accelerator pedal position 32 depends. A second calculation unit 35 calculated advantageously depending on the theoretical torque 34 and at least a first dead time 36 a delayed torque 37 the drive component 2nd , hereinafter also referred to as M _{VM, i} . The first dead time 36 can be caused, for example, by a delay in digital data processing in the control unit 6 caused.

A third calculation unit 38 preferably calculated depending on the target torque 29 M _{Dyno, s} and a second dead time 39 a second delayed torque 40 , hereinafter also referred to as M _{Dyno, i} . The second dead time 39 can, for example, by process times of the frequency converter 9 , in particular by setting the control currents. A fourth calculation unit advantageously adds 41 the delayed torque 37 the drive component 2nd with the second delayed torque 40 to a total torque 42 . The total torque 42 can be understood as a sum of all torques on the shaft 4th Act. Depending on the total torque 42 advantageously calculates a fourth calculation unit 43 taking into account a moment of inertia J _{VM of} the drive component 2nd and a moment of inertia J _{VM of} the electric motor 5 a theoretical angular acceleration 44 the wave 4th .

A fifth calculation unit 45 preferably determined based on the theoretical angular acceleration 44 , preferably by performing an integration of the theoretical angular acceleration 44 , a change in a theoretical angular velocity 46 the wave 4th . Furthermore, a sixth calculation unit advantageously calculates 47 based on the change in the theoretical angular velocity 46 the wave 4th and a third dead time 48 a change in angular velocity 49 the wave 4th . The third dead time 48 can be caused by a time delay in communication between a speed sensor 13 for measuring the speed of the shaft 4th and the control unit 6 caused. The control unit is advantageous 6 set up by changing the angular velocity 49 the wave 4th the future value of the shaft speed 4th to calculate for the future period. For example, the control device 6 the future value of the shaft speed 4th by adding the calculated change in angular velocity 49 the wave 4th to one with the speed sensor 13 measured value of the speed of the shaft 4th to calculate. In this case, the speed of the shaft 4th the state quantity of the wave 4th .

The kinematic system 8th can as one in 4th shown controlled system 14 of a control loop 20th be considered using a regulator 15 of the control unit 6 is adjustable. The regulator 15 shows the kinematic model 7 , for example in the form of the calculation module 30th , and is preferably set up as a function of a command variable 16 and a first value of the state variable 17th the wave 4th a first value of a manipulated variable 18th to calculate. The manipulated variable 18th and a disturbance 19th can on the controlled system 14 act by changing the respective manipulated variable 18th and the disturbance variable 19th a change in a value of the state quantity 17th the wave 4th cause. The state variable 17th the wave 4th is in a control variable of the control loop 20th . An embodiment is to be considered below in which the setpoint torque, M _{Dyno, s} , is the manipulated variable 18th , which is also referenced below with u, and the target speed n _s the reference variable 16 is. In the first possible embodiment of the kinematic model 7 the controller calculates 15 preferably a value of the manipulated variable 18th at least depending on the future value of the shaft speed 4th . This takes the controller into account 15 an estimated future behavior of the kinematic system 8th .

wobei ω die Drehzahl der Welle 4, M̈_VM,i eine zweifache Ableitung des verzögerten Drehmomentes 37 M̈_VM,i der Antriebskomponente 2 nach der Zeit und d einen Wert der Störgröße 19 des kinematischen Systems 8 bezeichnet. Das verzögerte Drehmoment 37 M̈_VM,i der Antriebskomponente 2, wie es oben beschrieben anhand des Kennfeldes 33 ermittelbar ist, kann gegenüber einem realen Drehmoment, das von der Antriebskomponente 2 auf die Welle 4 wirkt, eine Abweichung aufweisen. Da sowohl bei der ersten als auch bei der zweiten Ausführungsform des kinetischen Modells 7 das verzögerte Drehmoment 37 integrierend auf die Regelgröße wirkt, kann sich diese Abweichung zu einem stationären Regelfehler ausbilden. Dadurch, dass die Störgröße d ein Element des Zustandsvektors x ist, kann eine Kompensation dieses stationären Regelfehlers erleichtert werden.The following is a second variant of the kinematic model 7 to be discribed. The kinematic model 7 is in this variant in the form of a software module of the control unit 6 to calculate the future value of the state variable of the shaft. Preferably, at least the relationship described in equation (1) applies to this calculation. The state variable 17th the wave 4th is used in this variant as one of several state variables of the kinematic system 8th processed using the control unit. The control unit 6 is preferably set up a state of the kinematic system 8th using a state vector x, which is the state quantity of the kinematic system 8th summarizes, to describe, in particular for calculating the future value of the state variable of the wave 4th in a volatile memory of a CPU 21 of the control unit 6 such as a cache: $$x={\left[\begin{array}{cccc}\begin{array}{ccc}{x}_1& x_{\mathrm{2nd}}& x_{\mathrm{3rd}}\end{array}& x_{\mathrm{4th}}& x_5& x_6\end{array}\right]}^T={\left[\begin{array}{cccc}\begin{array}{ccc}\omega & M_{DynO,i}& M_{VM,i}\end{array}& {\ddot{M}}_{VM,i}& M_{VM,s}& d\end{array}\right]}^T,$$

where ω is the speed of the shaft 4th , M̈ _{VM, i} a double derivative of the decelerated torque 37 M̈ _{VM, i} the drive component 2nd according to time and d a value of the disturbance variable 19th of the kinematic system 8th designated. The delayed torque 37 M̈ _{VM, i} the drive component 2nd , as described above using the map 33 can be determined in relation to a real torque generated by the drive component 2nd on the wave 4th acts, have a deviation. Because both in the first and in the second embodiment of the kinetic model 7 the delayed torque 37 has an integrating effect on the controlled variable, this deviation from a stationary control error can develop. The fact that the disturbance variable d is an element of the state vector x makes it easier to compensate for this stationary control error.

The entries of the state vector x assume different values at different times. In the following, values of the state vector x which relate to a first point in time k are referenced with k and values of the state vector x which relate to a second point in time after the first point in time, i.e. future, refer to time k + 1, referenced with k + 1. In the same way, values of the state vector x which relate to a future point in time k + n, which is temporally seen n time steps behind the first time step k, are referenced with k + n. Accordingly, x (k) means the state vector with its values at the first point in time k and x (k + 1) means the state vector with its values at the second point in time k + 1. An analog referencing using the terms k, k + 1, k + n applies to the manipulated variable u.

$$\begin{array}{l}y\left(k\right)=Cx\left(k\right),\\ x\in {\Re }^n,u\in {\Re }^m,y\in {\Re }^r,A\in {\Re }^{n\times n},B\in {\Re }^{n\times m},C\in {\Re }^{r\times n},mit\\ A=\left[\begin{array}{cccccc}1& \frac{T_A}{J_{VM}+J_{Dyno}}& \frac{T_A}{J_{VM}+J_{Dyno}}& 0& 0& 0\\ 0& 1-\frac{T_A}{T_D}& 0& 0& 0& 0\\ 0& 0& 1& 0& 0& 0\\ 0& 0& -1+\frac{2D{T}_A}{T_{VM}}-\frac{T_A^2}{T_{VM}^2}& 2-\frac{2D{T}_A}{T_{VM}}& \frac{T_A^2}{T_{VM}^2}& 1\\ 0& 0& 0& 0& 1& 0\\ 0& 0& 0& 0& 0& 1\end{array}\right],\\ B=\left[\begin{array}{c}0\\ \frac{T_A}{T_D}\\ 0\\ 0\\ 0\\ 0\end{array}\right],\text{ und }{C}^T=\left[\begin{array}{c}1\\ 0\\ 0\\ 0\\ 0\\ 1\end{array}\right],\end{array}$$

wobei y(k) einen gemessenen Wert der Drehzahl der Welle 4, d.h. der Regelgröße 17, zum ersten Zeitpunkt beschreibt und D eine Dämpfungskonstante, T_VM eine Zeitkonstante der Antriebskomponente 2 und T_D eine Zeitkonstante des Elektromotors 5 beschreiben. Ein Wert der Regelgröße 17 für die unterschiedlichen Zeitpunkte k, k+1, k+n ist demnach durch einen jeweiligen Term y(k), y(k+1), y(k+n) beschrieben. Die Dämpfungskonstante D ist vorzugsweise als eine Konstante eines Tiefpassfilters zweiter Ordnung ausgebildet. Die Dämpfungskonstante D kann durch eine Untersuchung eines Zusammenhangs zwischen einer ersten Ableitung des verzögerten Drehmomentes 37 der Antriebskomponente 2 und der zweiten Ableitung des verzögerten Drehmomentes der Antriebskomponente 2 gewonnen werden. Weiterhin kann die Zeitkonstante T_D des Elektromotors 5 durch eine Untersuchung eines Zusammenhanges zwischen einer ersten Ableitung des verzögerten zweiten Drehmomentes 40 und dem zweiten verzögerten Drehmoment 40 gewonnen werden.The control unit 6 is preferably set up and designed, preferably independently of that in 3rd shown calculation module 30th To calculate values of the state vector x (k + 1) as a function of values of the state vector x (k) and a value of the manipulated variable u (k), preferably according to equation (1). The values of the state vector x (k + 1) can be used as future values of the state variables of the kinematic system 8th be understood. Accordingly, the first entry of the state vector x (k + 1) denotes the future value of the state variable 17th the wave 4th . $$x\left(k+1\right)=Ax\left(k\right)+Bu\left(k\right),$$

$$\begin{array}{l}y\left(k\right)=\mathrm{C.}x\left(k\right),\\ x\in {\Re }^n,u\in {\Re }^m,y\in {\Re }^r,A\in {\Re }^{n\times n},B\in {\Re }^{n\times m},\mathrm{C.}\in {\Re }^{r\times n},mit\\ A=\left[\begin{array}{cccccc}1& \frac{T_A}{J_{VM}+J_{DynO}}& \frac{T_A}{J_{VM}+J_{DynO}}& 0& 0& 0\\ 0& 1-\frac{T_A}{T_D}& 0& 0& 0& 0\\ 0& 0& 1& 0& 0& 0\\ 0& 0& -1+\frac{\mathrm{2nd}D{T}_A}{T_{VM}}-\frac{T_A^{\mathrm{2nd}}}{T_{VM}^{\mathrm{2nd}}}& \mathrm{2nd}-\frac{\mathrm{2nd}D{T}_A}{T_{VM}}& \frac{T_A^{\mathrm{2nd}}}{T_{VM}^{\mathrm{2nd}}}& 1\\ 0& 0& 0& 0& 1& 0\\ 0& 0& 0& 0& 0& 1\end{array}\right],\\ B=\left[\begin{array}{c}0\\ \frac{T_A}{T_D}\\ 0\\ 0\\ 0\\ 0\end{array}\right],\text{and}{\mathrm{C.}}^T=\left[\begin{array}{c}1\\ 0\\ 0\\ 0\\ 0\\ 1\end{array}\right],\end{array}$$

where y (k) is a measured value of the speed of the shaft 4th , ie the controlled variable 17th , describes at the first point in time and D a damping constant, T _VM a time constant of the drive component 2nd and T _{D is} a time constant of the electric motor 5 describe. A value of the controlled variable 17th for the different times k, k + 1, k + n is accordingly described by a respective term y (k), y (k + 1), y (k + n). The damping constant D is preferably designed as a constant of a second order low-pass filter. The damping constant D can be determined by examining a relationship between a first derivative of the decelerated torque 37 the drive component 2nd and the second derivative of the decelerated torque of the drive component 2nd be won. Furthermore, the time constant T _{D of} the electric motor 5 by examining a relationship between a first derivative of the delayed second torque 40 and the second delayed torque 40 be won.

With T _A is a time step size of the control unit 6 designated with which the state vector x with the help of the control unit 6 is updated. An update of the state vector x means a calculation of the temporally next state vector x (k + 1) as a function of the state vector x (k) according to equation (1). Equation (1) and equation (2) together describe discrete-time state space equations for the kinematic system 8th .

The control unit advantageously determines 6 the values of the state variables ω, M _{Dyno, i} , M _{VM, i} , M _{VM, i, 2} , M _{VM, s} , d as a function of the first dead time described above 36 , second dead time 39 and / or third dead time 48 . The dead times 36 , 39 , 48 can the control unit 6 for example by taking into account the control unit 6 for each of the dead times 36 , 39 , 48 m processed further equations of state analogously to equation (1), where m is an integer quotient with the corresponding dead time 36 , 39 respectively. 48 as a dividend and the sampling time as a divisor.

wobei N_p einen Prädiktionshorizont und N_u einen Stellhorizont bezeichnet. Der Regler 15 führt bevorzugt eine Veränderung der Stellgröße u bis einschließlich zum Stellhorizont N_u in Abhängigkeit einer mit dem Steuergerät 6 berechneten Stellgrößendifferenz Δu für die Zeitpunkte k bis k+N_u durch. Bei den zeitlich nach dem Stellhorizont N_u liegenden Zeitpunkten hält der Regler 15 die Stellgröße bevorzugt konstant. Die Veränderung der Stellgröße u kann das Steuergerät 6 bevorzugt nach Gleichung (6) berechnen: $$u\left(k\right)=u\left(k-1\right)+\mathrm{\Delta }u\left(k\right)$$

In an advantageous embodiment, the control device 6 set up and designed to calculate several further future values of the state vector and the manipulated variable for several further future times k + 2 ... k + N _u ... k + N _p . For example, the control unit can perform calculations of these further future values according to equations (3), (4) and (5): $$\begin{array}{c}\text{equation}\left(\mathrm{3rd}\right):\text{x}\left(k+\mathrm{2nd}|k\right)=Ax\left(k+1|k\right)+Bu\left(k|k\right)+B\Delta u\left(k+1|k\right)\\ ⋮\\ \text{equation}\left(\mathrm{4th}\right):\text{x}\left(k+N_u|k\right)=Ax\left(k+N_u-1|k\right)+Bu\left(k+N_u-\mathrm{2nd}|k\right)+B\Delta u\left(k+N_{u}1|k\right)\\ ⋮\\ \text{equation}\left(5\right):\text{x}\left(k+N_p|k\right)=Ax\left(k+N_p-1|k\right)+Bu\left(k+N_u-1|k\right)\end{array}$$

where N _p denotes a prediction horizon and N _u a positioning horizon. The regulator 15 preferably changes the manipulated variable u up to and including the actuating horizon N _u as a function of one with the control unit 6 calculated manipulated variable difference Δu for the times k to k + N _u . The controller stops at the times lying after the actuating horizon N _u 15 the manipulated variable is preferably constant. The control unit can change the manipulated variable u 6 preferably calculate according to equation (6): $$u\left(k\right)=u\left(k-1\right)+\mathrm{\Delta }u\left(k\right)$$

Intermediate steps can be done using the control unit 6 can be calculated analogously to equations (3), (4) and (5). A prediction of the controlled variable 17th can the control unit 6 advantageously determine according to equation (7): $$\begin{array}{l}y\left(\cdot |k\right)=\left[\begin{array}{c}\mathrm{C.}x\left(k+1|k\right)\\ \mathrm{C.}x\left(k+\mathrm{2nd}|k\right)\\ ⋮\\ \mathrm{C.}x\left(k+N_u|k\right)\\ ⋮\\ \mathrm{C.}x\left(k+N_p|k\right)\end{array}\right]=\left[\begin{array}{c}\mathrm{C.}{A}^1\\ \mathrm{C.}{A}^{\mathrm{2nd}}\\ ⋮\\ \mathrm{C.}{A}^{N_u}\\ ⋮\\ \mathrm{C.}{A}^{N_p}\end{array}\right]x\left(k\right)+\left[\begin{array}{c}\mathrm{C.}B\\ \mathrm{C.}\left(A+\mathrm{I.}\right)B\\ ⋮\\ \mathrm{C.}\left(A^{N_u-1}+A^{N_u-\mathrm{2nd}}+\cdots +\mathrm{I.}\right)B\\ ⋮\\ \mathrm{C.}\left(A^{N_p-1}+A^{N_p-\mathrm{2nd}}+\cdots +\mathrm{I.}\right)B\end{array}\right]u\left(k-1\right)\\ +\left[\begin{array}{cccc}\mathrm{C.}B& 0& \cdots & 0\\ \mathrm{C.}\left(A+\mathrm{I.}\right)B& \mathrm{C.}B& \cdots & 0\\ ⋮& ⋮& \ddots & ⋮\\ \mathrm{C.}\left(A^{N_u-1}+\cdots +\mathrm{I.}\right)B& \mathrm{C.}\left(A^{N_u-\mathrm{2nd}}+\cdots +\mathrm{I.}\right)B& \cdots & \mathrm{C.}B\\ ⋮& ⋮& \ddots & ⋮\\ \mathrm{C.}\left(A^{N_p-1}+\cdots +\mathrm{I.}\right)B& \mathrm{C.}\left(A^{N_p-\mathrm{2nd}}+\cdots +\mathrm{I.}\right)B& \cdots & \mathrm{C.}\left(A^{N_p-N_u}+\cdots +\mathrm{I.}\right)B\end{array}\right]\left[\begin{array}{c}\Delta u\left(k|k\right)\\ \Delta u\left(k+1|k\right)\\ ⋮\\ \Delta u\left(k+N_u-1|k\right)\end{array}\right]\end{array}$$

Equation (7) can be iteratively substituted by state space equations (1), (2) in equation (3) to predict the state vector x (k + 2 | k) following the state vector x (k + 1 | k). to be created.

wobei der linke und konstante Teil Kx(k) + L_MPRu(k - 1) der rechten Seite der Gleichung (8) zu einem Term f(k) zusammengefasst werden kann. Eine Verwendung der Gleichung (8) kann es erleichtern, einen Differenzstellgrößenvektor Δu(· |k) zu optimieren. Eine Differenz zwischen der Führungsgröße 16, die im Folgenden auch mit w(. |k) bezeichnet wird, und f(k) kann das Steuergerät 6 bevorzugt in Form eines zum Zeitpunkt k konstanten Regelfehlers e(·|k), wie in Gleichung (9) beschrieben, verarbeiten: $$e\left(\cdot |k\right)=w\left(\cdot |k\right)-f\left(k\right)$$

The control device advantageously has 6 an implementation of the system of equations (7) in a simplified form according to equation (8), in which the parameter matrices A, B, C are substituted by the matrices K, L, M: $$\begin{array}{l}y\left(\cdot |k\right)=\stackrel{f\left(x\right)}{\overbrace{Kx\left(k\right)+L_{MPR}u\left(k-1\right)}}+M\Delta u\left(\cdot |k\right),\\ K\in {\Re }^{r\cdot N_p\times m},L_{MPR}\in {\Re }^{r\cdot N_p\times m},M\in {\Re }^{r\cdot N_p\times m\cdot N_u},\end{array}$$

whereby the left and constant part Kx (k) + L _MPR u (k - 1) of the right side of the equation (8) can be combined into a term f (k). Using equation (8) can make it easier to optimize a differential manipulated variable vector Δu (· | k). A difference between the benchmark 16 which are below is also denoted by w (. | k), and f (k) can the control unit 6 Process preferably in the form of a control error e (· | k) which is constant at time k, as described in equation (9): $$e\left(\cdot |k\right)=w\left(\cdot |k\right)-f\left(k\right)$$

In an advantageous embodiment, the control device 6 trained and set up to calculate a value of a cost function J (Δu (· Ik)) according to equation (10) depending on the control error e (· | k): $$\begin{array}{l}J\left(\Delta u\left(\cdot |k\right)\right)={\left(e\left(\cdot |k\right)+M\Delta u\left(\cdot |k\right)\right)}^{T}Q(e\left(\cdot |k\right)+M\Delta u\left(\cdot |k\right)+\Delta u^T\left(\cdot |k\right)R_u{\Delta }_u\left(\cdot |k\right)\\ =\Delta u^T\left(\cdot |k\right)\left(M^{T}QM+R_u\right)\Delta u\left(\cdot |k\right)+\mathrm{2nd}\Delta u^T\left(\cdot |k\right)M^{T}Qe\left(\cdot |k\right)+e^T\left(\cdot |k\right)Qe\left(\cdot |k\right),\text{and}\\ Q\in {\Re }^{r\cdot N_p\times r\cdot N_p},R_u\in {\Re }^{m\cdot N_u\times m\cdot N_u},e\in {\Re }^{r\cdot N_p\times 1}.\end{array}$$

Furthermore, the control unit 6 preferably set up to minimize the value of the cost function J (Δu (· Ik)) using an optimization method in order to calculate a change in the manipulated variable Δu for the times k to k + N _u . The control unit 6 preferably sends a signal to adjust the manipulated variable 18th to the frequency converter 9 . The control unit 6 generates the signal for adjusting the manipulated variable 18th preferably by transforming the value Δu (k) for the first time k determined with the aid of the cost function to a digital value for controlling the frequency converter 9 . After adjusting the manipulated variable 18th the control unit forms at time k 6 prefers equations (1) to (10) again, whereby all values that were previously referenced with k + n + 1 are referenced with k + n, where n is an integer between 0 and N _p . Then the control unit leads 6 prefers the optimization process again to determine the value of the cost function J (Δu (· | k)) and to calculate a new value Δu (k) for the first time k.

In that the control unit 6 The control unit calculates the future values of the state vector for the times k ... k + N _p 6 also future values of the disturbance variable 19th , which is denoted by d, for the times k ... k + N _p , because the state vector is the disturbance variable 19th having.

$$b_{u}Z\le Z\Delta u\le b_{0}Z,$$

wobei H eine Hessematrix, g einen Gradientenvektor, Z eine Beschränkungsmatrix, b_u eine untere Beschränkung und b₀ eine obere Beschränkung bezeichnen.The control device is advantageously set up, the manipulated variable 18th to limit downwards and / or upwards. The control unit leads to this 18th preferably minimizing the term ( 11 ) taking into account the boundary conditions expressed by inequality (12) by: $$\frac{1}{\mathrm{2nd}}\Delta u^{T}H\Delta \text{u}+\Delta {\text{u}}^{T}G$$

$$b_u\mathrm{Z.}\le \mathrm{Z.}\Delta u\le b_0\mathrm{Z.},$$

where H denotes a Hessian matrix, g a gradient vector, Z a restriction matrix, b _u a lower restriction and b ₀ an upper restriction.

The control unit is used to determine the values of the state vector x (k) at the first point in time k 6 preferably set up to estimate the values of the state vector x (k) in one prediction step and using at least one measured value of one of the state variables of the kinematic system represented by the state vector x 8th to correct in one correction step.

$$\begin{array}{l}\widehat{x}\left(k\right)=\widehat{x}{\left(k\right)}^{S1}+L\left(y\left(k\right)-C\widehat{x}{\left(\text{k}\right)}^{S1}\right),\text{ mit}\\ L=\frac{P{\left(k\right)}^{S1}{C}^T}{CP{\left(k\right)}^{S1}{C}^T+R_L},\end{array}$$

wobei R_L eine konstante Kovarianzmatrix von Messgrößen, die zum Beispiel die gemessene Drehzahl der Welle 4 umfassen, bezeichnet. Prinzipiell ist es möglich, dass neben der Drehzahl auch weitere Zustandsgrößen des kinematischen Systems 8 gemessen werden und deren Messwerte in Form von Einträgen des Vektors y(k) bei dem Korrekturschritt mit einfließen. Die Matrix L ist eine Beobachtermatrix. Da die Gleichung zumindest einen Messwert, insbesondere den gemessenen Wert der Drehzahl der Welle 4, enthält, wird die Gleichung (14) im Folgenden Beobachtergleichung genannt. Bei einer Durchführung des Prädiktionsschrittes berechnet das Steuergerät 6 bevorzugt eine geschätzte Kovarianzmatrix P(k)^S1 nach Gleichung (15): $$P{\left(k\right)}^{S1}=AP\left(k-1\right)A^T+Q_L,$$

wobei Q_L eine konstante Kovarianzmatrix der Zustandsgrößen, die durch die Elemente des Zustandsvektors x repräsentiert werden, bezeichnet.The control unit is advantageous 6 set up to carry out the prediction step according to equation (13) and the correction step according to equation (14), where x̂ (k) ^S1 denotes an estimated state vector and x̂ (k) denotes a corrected estimated state vector: $$\widehat{x}{\left(k\right)}^{S1}=A\widehat{x}\left(k-1\right)+Bu\left(k-1\right)$$

$$\begin{array}{l}\widehat{x}\left(k\right)=\widehat{x}{\left(k\right)}^{S1}+L\left(y\left(k\right)-\mathrm{C.}\widehat{x}{\left(\text{k}\right)}^{S1}\right),\text{With}\\ L=\frac{P{\left(k\right)}^{S1}{\mathrm{C.}}^T}{\mathrm{C.}P{\left(k\right)}^{S1}{\mathrm{C.}}^T+R_L},\end{array}$$

where R _{L is} a constant covariance matrix of measured variables, for example the measured speed of the shaft 4th include, referred to. In principle, it is possible that in addition to the speed, other state variables of the kinematic system 8th are measured and their measured values in the form of entries of the vector y (k) are included in the correction step. The matrix L is an observer matrix. Since the equation has at least one measured value, in particular the measured value of the speed of the shaft 4th , contains, equation (14) is called the observer equation below. The control unit calculates when the prediction step is carried out 6 preferably an estimated covariance matrix P (k) ^S1 according to equation (15): $$P{\left(k\right)}^{S1}=AP\left(k-1\right)A^T+Q_L,$$

where Q _L denotes a constant covariance matrix of the state variables represented by the elements of the state vector x.

The control unit calculates when the correction step is carried out 6 preferably a corrected estimated covariance matrix P (k) according to equation (16): $$P\left(k\right)=\left(\mathrm{I.}-L\mathrm{C.}\right)P{\left(k\right)}^{S1}$$

If a corrected estimated covariance matrix P (k) is not yet available when the prediction step is carried out for the first time, an initial covariance matrix P _{0 is} used instead of the corrected estimated covariance matrix P (k). Accordingly, an initial corrected estimated state vector x _{0 can be used} when the prediction step is carried out for the first time.

When the state vector x (k) is calculated in accordance with equations (13) and (14), the entries of the state vector x are regarded as random variables which fluctuate normally distributed around individual entries of the estimated state vector x̂ (k). Therefore, the state vector x can be described by the corrected estimated state vector x̂ (k) and the corrected estimated covariance matrix P (k). Analogously, the state vector x can be described by the estimated state vector x̂ (k) ^S1 and the estimated covariance matrix P (k) ^S1 .

The control unit advantageously sets 6 the state vector x (k) for calculations according to equations (1) to (10) equal to the corrected estimated state vector x̂ (k). This has the advantage that information about the measured value of the speed y (k) of the shaft 4th if the state vector x is used further, such as when determining the change in the manipulated variable Δu (k) according to equations (1) to (10), can be taken into account. The measured value of the speed y (k) of the shaft 4th flows in a calculation of the corrected estimated state vector x) (k) according to equation (14) using the control unit 6 with a.

wobei d(k) Werte von Störgrößen, darunter die Störgröße d, zum Zeitpunkt k bezeichnen und η(k) Rauschterme der Werte der Störgrößen d(k) angeben und die Matrizen A_xd und A_d jeweils eine Störgrößendynamik der Störgrößen in Bezug auf das kinematische System 8 angeben. Mit x_E(k + 1) ist der um die Werte d(k + 1) erweiterte Zustandsvektor x(k + 1) zum Zeitpunkt k + 1 bezeichnet. Um darüber hinaus den Einfluss der Störgröße d auf das kinematische System berücksichtigen zu können, führt das Steuergerät den Korrekturschritt bevorzugt gemäß Gleichung (18) anstatt nach Gleichung (14) aus: $${\widehat{x}}_E\left(k+1\right)=\left[\begin{array}{cc}A& A_{xd}\\ 0& A_d\end{array}\right]\left[\begin{array}{l}\widehat{x}\left(k\right)\\ \widehat{d}\left(k\right)\end{array}\right]+\left[\begin{array}{l}{L}_x\\ L_d\end{array}\right]\left(y\left(k\right)-C\widehat{x}\left(k\right)-C_d\widehat{d}\left(k\right)\right),$$

wobei d(k) geschätzte Werte der Störgrößen bezeichnen, die auch einen geschätzten Wert der Störgröße 19 umfassen. Möglich ist, dass die Störgröße 19, d.h. die Störgröße d die einzige durch Gleichung (18) berücksichtigte Störgröße ist. Jedoch drückt die Verwendung des Vektors d(k) und d(k) aus, dass bei dem Prädiktionsschritt und dem Korrekturschritt auch mehrere Störgrößen berücksichtigt werden können. Die Matrix C_d gibt einen Zusammenhang zwischen der Messgröße y(k) und den Störgrößen, insbesondere der Störgröße d, an. Mit x̂_E(k+1) ist der korrigierte geschätzte erweiterte Zustandsvektor zum Zeitpunkt k + 1 bezeichnet.In an advantageous embodiment, the control device 6 set up and designed to influence the implementation of the correction of the estimated state vector x̂ (k) ^{S1 of} the kinematic system Disturbance variable d to be considered on the kinematic system. For this purpose, the control unit preferably executes the prediction step according to equation (17) instead of according to equation (13): $$x_E\left(k+1\right)=\left[\begin{array}{l}x\left(k+1\right)\\ d\left(k+1\right)\end{array}\right]=\left[\begin{array}{cc}A& A_{xd}\\ 0& A_d\end{array}\right]\left[\begin{array}{l}x\left(k\right)\\ d\left(k\right)\end{array}\right]+\left[\begin{array}{l}B\\ 0\end{array}\right]u\left(k\right)+\eta \left(k\right),$$

where d (k) denote values of disturbance variables, including disturbance variable d, at time k and η (k) indicate noise terms of the values of disturbance variables d (k) and the matrices A _xd and A _d each have a disturbance variable dynamics of the disturbance variables in relation to the kinematic system 8th specify. X _E (k + 1) denotes the state vector x (k + 1) expanded by the values d (k + 1) at time k + 1. In order to be able to take into account the influence of the disturbance variable d on the kinematic system, the control unit preferably carries out the correction step according to equation (18) instead of equation (14): $${\widehat{x}}_E\left(k+1\right)=\left[\begin{array}{cc}A& A_{xd}\\ 0& A_d\end{array}\right]\left[\begin{array}{l}\widehat{x}\left(k\right)\\ \widehat{d}\left(k\right)\end{array}\right]+\left[\begin{array}{l}{L}_x\\ L_d\end{array}\right]\left(y\left(k\right)-\mathrm{C.}\widehat{x}\left(k\right)-{\mathrm{C.}}_d\widehat{d}\left(k\right)\right),$$

where d (k) denote estimated values of the disturbance variables, which also an estimated value of the disturbance variable 19th include. It is possible that the disturbance variable 19th , ie the disturbance variable d is the only disturbance variable taken into account by equation (18). However, the use of the vector d (k) and d (k) expresses that a plurality of disturbance variables can also be taken into account in the prediction step and the correction step. The matrix C _d indicates a relationship between the measured variable y (k) and the disturbance variables, in particular the disturbance variable d. X̂ _E (k + 1) denotes the corrected estimated extended state vector at time k + 1.

In a further embodiment, the control device 6 set up and designed to carry out the prediction step according to equation (17) without the correction step according to equation (18). This embodiment can be advantageous with an education of the kinematic model 7 can be combined according to the first variant described above. In this case, it is possible to take into account the influence of the disturbance variable d as part of a pre-control.

The control unit 6 preferably has a software module 22 which provides an implementation of the relationships described in equations (1) to (18). The implementation can be carried out using matrices, vectors and arithmetic operations for the matrices and vectors. The control unit provides the implementation of the influence of the disturbance variable on the kinematic system, which is expressed by equation (18) 6 preferably a disturbance observer ready. By implementing the prediction and correction step according to equations (13) and (17) or (14) and (18), the control unit 6 advantageously a Kalman filter is ready. This makes it possible to calculate the respective prediction and correction step as a respective linear combination of previous prediction and correction steps. Such a recursive calculation to estimate the state vector x (k) can reduce the computing power required.

With the test bench described so far 1 can use a following procedure to test the drive component 2nd for the vehicle 3rd be performed. The procedure has the in 5 shown steps. In a first step 51 becomes the future value of the state quantity of the wave 4th for the future period using the kinematic model 7 for modeling the kinematic system 8th calculated. The kinematic system 8th has at least the drive component 2nd , the electric motor 5 and the wave 4th on. In a second step, the state variable of the wave 4th depending on the future value of the state variable of the shaft 4th regulated. A regulation of the state variable of the wave 4th is preferred with the control unit 6 , as it can be designed according to one of the variants described above.

6 shows a drive system 60 for another vehicle 61 . The drive system 60 has another drive component 62 , another wave 63 , another electric motor 64 with the other drive component 62 about the further wave 63 is mechanically coupled, and another control unit 65 with another kinematic model 66 for modeling another kinematic system 67 . The other kinematic system 67 has at least the further drive component 62 , the other electric motor 64 and the further wave 63 on. The other control unit 65 of the drive system 60 is set up and trained using the other kinematic model 66 at least one future value of a state variable of the further wave 63 to calculate for a future period and depending on the future value of the state variable of the further wave 63 the state quantity of the further wave 63 to regulate. The other drive component 62 can like the drive component described above 2nd be trained. The further kinematic model is preferred 66 like the kinematic model described above 66 executed. The other control unit 65 preferably has at least the same functions as the control device described above 6 how it can be designed according to one of the above-mentioned variants.

QUOTES INCLUDE IN THE DESCRIPTION

This list of documents listed by the applicant has been generated automatically and is only included for better information for the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• WO 2011038429 A1 [0002]

Claims (10)

Test bench (1) for testing a drive component (2) for a vehicle (3), the test bench (1) has a shaft (4), an electric motor (5) which mechanically connects with the drive component (2) via the shaft (4) and a control device (6) with a kinematic model (7) for modeling a kinematic system (8), which has at least the drive component (2), the electric motor (5) and the shaft (4), the control device ( 6) is set up and designed to use the kinematic model (7) to calculate at least one future value of a state variable of the shaft (4) for a future period and, depending on the future value of the state variable of the shaft (4), the state variable of the shaft (4 ) to regulate. Test bench (1) Claim 1 The control device (6) is set up and designed to calculate at least one future value of a disturbance variable (19) of the kinematic system (8) for the future period and, depending on the future value of the disturbance variable (19), the state variable of the shaft (4 ) to regulate. Test bench (1) Claim 1 or 2nd , wherein the state variable of the shaft (4) is a speed of the shaft (4). Test stand (1) according to one of the preceding claims, wherein the test stand (1) has a sensor (13) for detecting a measured value of a state variable of the kinematic system (8) and the control device (6) is set up and designed, a value of the state variable of the to estimate the kinematic system (8) and use the measured value of the state variable of the kinematic system (8) to correct the estimated value of the state variable of the kinematic system (8) and as a result of the correction a corrected estimated value of the state variable of the kinematic system (8) to determine and to calculate the future value of the state variable of the shaft (4) as a function of the corrected estimated value of the state variable of the kinematic system (8). Test bench (1) Claim 2 and 4th The control device (6) is set up and designed to take into account an influence of the disturbance variable (19) on the kinematic system when carrying out the correction of the estimated value of the state variable of the kinematic system (8). Test stand (1) according to one of the preceding claims, wherein the control device (6) is set up and designed to calculate the future value of the state variable of the shaft (4) as a function of at least one dead time (36; 39; 48) of the test stand (1) . Test stand (1) according to one of the preceding claims, wherein the control device (6) is set up and designed to determine the future value of the state variable of the shaft (4) as a function of an estimated torque (34) of the drive component (2) and the estimated torque (34) of the drive component (2) using a characteristic diagram of the drive component (2). Drive system (60) for a vehicle (61), the drive system (60) having a drive component (62), a shaft (63), an electric motor (64) which is mechanically coupled to the drive component (62) via the shaft (63) and a control device (65) with a kinematic model (66) for modeling a kinematic system (67) which has at least the drive component (62), the electric motor (64) and the shaft (63), the control device (65 ) is set up and designed to use the kinematic model (66) to calculate at least one future value of a state variable of the shaft (63) for a future period and, depending on the future value of the state variable of the shaft (63), the state variable of the shaft (63) to regulate. Method for testing a drive component (2) for a vehicle (3), the drive component (2) being mechanically coupled to an electric motor (5) via a shaft (4), with the following steps: - Calculating a future value of a state variable of the shaft (4) for a future period using a kinematic model (7) for modeling a kinematic system (8) which at least the drive component (2), the electric motor (5) and the shaft (4 ) having, - Regulating the state variable of the shaft (4) depending on the future value of the state variable of the shaft (4). A computer program product comprising a program that, when executed by a computer, causes the computer to follow a method Claim 9 perform.

Images