Method for controlling a load machine during a test run with a drive train and test stand

Abstract

The present teaching includes a speed control of a side shaft of a drive train connected to a dynamometer on a drive train test stand in which a torque (M.sub.Fxi) caused by the longitudinal force (F.sub.Xi) calculated in a simulation model is additionally transferred to the control unit, and from this a compensation torque (M.sub.Ki) is calculated in the control unit as a function of the longitudinal force (F.sub.Xi) caused by the torque (M.sub.Fxi) and a deviation (A.sub.Ji) between a moment of inertia (J.sub.Bi) of the dynamometer and a moment of inertia (J.sub.Ri) of the simulated vehicle wheel, the control unit calculates a torque (M.sub.REi,) from the setpoint speed (n.sub.Bi,set) with a speed controller and a torque (M.sub.Bi,soll) to be set with the dynamometer is calculated as the sum of the compensation torque (M.sub.Ki) and the torque (M.sub.REi) calculated by the speed controller and set by the dynamometer.

Claims

1. A method for carrying out a test run on a drive train test stand on which a drive train with at least one side shaft is arranged and this side shaft is connected to a dynamometer and the speed of the side shaft is regulated in a control unit, wherein in a simulation model a longitudinal force of a tire of a vehicle wheel, represented as F.sub.Xi, is simulated with the simulation model, and a setpoint speed of the dynamometer to be set is calculated, and the setpoint speed is transferred to the control unit as a setpoint value of the speed control of the dynamometer, characterized in that a torque caused by the longitudinal force, represented by M.sub.Fxi, is calculated in the simulation model and is also transferred to the control unit, and therefrom a compensation torque, represented by M.sub.Ki, is calculated in the control unit as a function of the torque caused by the longitudinal force and a deviation between a moment of inertia of the dynamometer, represented by J.sub.Bi, and a moment of inertia of the simulated vehicle wheel, represented by J.sub.Ri, that the compensation torque is calculated from the relationship: $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+M_{\mathrm{Ri}}-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Bi}}\mathrm{or}$ $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+\left(1-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}\right)M_{\mathrm{Ri}},$ with a torque acting on the side shaft, represented by M.sub.Ri and a torque acting on the shaft of the dynamometer, represented by M.sub.Bi, or
M.sub.Ki=M.sub.Fxi+J.sub.Ri{dot over ()}.sub.RiJ.sub.Bi{dot over ()}.sub.Bi or M.sub.Ki=M.sub.Fxi+(J.sub.RiJ.sub.Bi){dot over ()}.sub.Ri, with an angular acceleration acting on the side shaft, represented by {dot over ()}.sub.Ri, and an angular acceleration acting on the shaft of the dynamometer, represented by {dot over ()}.sub.Bi, that the control unit calculates a control torque set with the dynamometer is calculated as the sum of the compensation torque and the control torque.

2. The method of claim 1, characterized in that the simulation model includes a wheel model having a tire model and a vehicle model, wherein the vehicle model calculates a longitudinal velocity the simulated vehicle and a vertical force of the vehicle wheel and transfers them to the wheel model and the wheel model having the tire model calculates the longitudinal force and transfers it to the vehicle model.

3. The method of claim 1, wherein the compensation torque is calculated using a rolling resistance torque of the tire of the vehicle wheel.

4. A drive train test stand with a drive train of a vehicle as a test specimen, wherein the drive train on the drive train test stand is subjected to a test run, wherein at least one side shaft of the drive train is connected with a dynamometer and a control unit is provided to control the speed of the side shaft in accordance with the specifications of the test run, and wherein a simulation model for simulating a vehicle wheel of the vehicle is implemented on the drive train test stand that calculates a longitudinal force of the tire of the vehicle wheel, represented as F.sub.Xi, and a setpoint speed of the dynamometer to be set, characterized in that in the control unit a compensation unit is provided that calculates a compensation torque, represented by M.sub.Ki, from a torque caused by the longitudinal force, represented by M.sub.Fxi, and a deviation between a moment of inertia of the dynamometer, represented by J.sub.Bi, and a moment of inertia of the simulated vehicle wheel, represented by J.sub.Ri, that the compensation torque is calculated from the relationship: $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+M_{\mathrm{Ri}}-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Bi}}\mathrm{or}$ $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+\left(1-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}\right)M_{\mathrm{Ri}},$ with a torque acting on the side shaft, represented by M.sub.Ri, and a torque acting on the shaft of the dynamometer, represented by M.sub.Bi, or
M.sub.Ki=M.sub.Fxi+J.sub.Ri{dot over ()}.sub.RiJ.sub.Bi{dot over ()}.sub.Bi or M.sub.Ki=M.sub.Fxi+(J.sub.RiJ.sub.Bi){dot over ()}.sub.Ri, with an angular acceleration acting on the side shaft, represented by {dot over ()}.sub.Ri, and an angular acceleration acting on the shaft of the dynamometer, represented by {dot over ()}.sub.Bi, that a speed controller is implemented in the control unit, which calculates a control torque from the setpoint speed, and that the dynamometer sets the sum of the compensation torque and the control torque on the drive train dynamometer.

5. The drive train test stand of claim 4, wherein the compensation torque is calculated using a rolling resistance torque of the tire of the vehicle wheel.

6. The drive train test stand of claim 4, wherein the simulation model includes a wheel model having a tire model and a vehicle model, wherein the vehicle model calculates a longitudinal velocity of the simulated vehicle and a vertical force of the vehicle wheel and transfers them to the wheel model and the wheel model having the tire model calculates the longitudinal force and transfers it to the vehicle model.

7. A drive train test stand, comprising: a drive train of a vehicle; a dynamometer connected with at least one side shaft of the drive train; a control unit controlling speed of the side shaft in accordance with specifications of a test run; a simulation model simulating a vehicle wheel of the vehicle that calculates a longitudinal force of a tire of the vehicle wheel, represented as F.sub.Xi, and a setpoint speed of the dynamometer; a compensation unit calculating a compensation torque, represented by M.sub.Ki, from a torque caused by the longitudinal force, represented by M.sub.Fxi, and a deviation between a moment of inertia of the dynamometer, represented by J.sub.Bi, and a moment of inertia of the simulated vehicle wheel, represented by J.sub.Ri; a speed controller calculating a control torque from the setpoint speed, and the dynamometer sets the sum of the compensation torque and the control torque; wherein the compensation torque is calculated from the relationship: $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+M_{\mathrm{Ri}}-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Bi}}\mathrm{or}$ $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+\left(1-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}\right)M_{\mathrm{Ri}},$ with a torque acting on the side shaft, represented by M.sub.Ri, and a torque acting on the shaft of the dynamometer, represented by M.sub.Bi; or
M.sub.Ki=M.sub.Fxi+J.sub.Ri{dot over ()}.sub.RiJ.sub.Bi{dot over ()}.sub.Bi or M.sub.Ki=M.sub.Fxi+(J.sub.RiJ.sub.Bi){dot over ()}.sub.Ri, with an angular acceleration acting on the side shaft, represented by {dot over ()}.sub.Ri, and an angular acceleration acting on the shaft of the dynamometer represented by {dot over ()}.sub.Bi.

8. The drive train test stand of claim 7, wherein the compensation torque is calculated using a rolling resistance torque of the tire of the vehicle wheel.

9. The drive train test stand of claim 7, wherein the simulation model includes a wheel model having a tire model and a vehicle model, wherein the vehicle model calculates a longitudinal velocity of the simulated vehicle and a vertical force of the vehicle wheel and transfers them to the wheel model and the wheel model having the tire model calculates the longitudinal force and transfers it to the vehicle model.

10. A method for carrying out a test run, comprising: arranging a drive train on a drive train test stand; connecting at least one side shaft of the drive train to a dynamometer; regulating a speed of the side shaft in a control unit; simulating in a simulation model a longitudinal force of a tire of a vehicle wheel, represented as F.sub.xi; calculating a setpoint speed of the dynamometer to be set; transferring the setpoint speed to the control unit as a setpoint value of the speed control of the dynamometer; calculating a torque caused by the longitudinal force, represented by M.sub.Fxi, in the simulation model; transferring the torque caused by the longitudinal force to the control unit; calculating a compensation torque, represented by M.sub.Ki, in the control unit as a function of the torque caused by the longitudinal force, represented by M.sub.Fxi, and a deviation between a moment of inertia of the dynamometer, represented by JB.sub.i, and a moment of inertia of the simulated vehicle wheel, represented by J.sub.Ri, setting a torque with the dynamometer calculated as a sum of the compensation torque and a control torque calculated from the setpoint speed with the speed controller; wherein the compensation torque is calculated from the relationship: $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+M_{\mathrm{Ri}}-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Bi}}\mathrm{or}$ $M_{\mathrm{Ki}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+\left(1-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}\right)M_{\mathrm{Ri}},$ with a torque acting on the side shaft, represented by M.sub.Ri, and a torque acting on the shaft of the dynamometer, represented by M.sub.Bi; or
M.sub.Ki=M.sub.Fxi+J.sub.Ri{dot over ()}.sub.RiJ.sub.Bi{dot over ()}.sub.Bi or M.sub.Ki=M.sub.Fxi+(J.sub.RiJ.sub.Bi){dot over ()}.sub.Ri, with an angular acceleration acting on the side shaft, represented by {dot over ()}.sub.Ri, and an angular acceleration acting on the shaft of the dynamometer represented by {dot over ()}.sub.Bi.

11. The method for carrying out a test run of claim 10, wherein the compensation torque is calculated using a rolling resistance torque of the tire of the vehicle wheel.

12. The method for carrying out a test run of claim 10, wherein the simulation model includes a wheel model having a tire model and a vehicle model, wherein the vehicle model calculates a longitudinal velocity of the simulated vehicle and a vertical force of the vehicle wheel and transfers them to the wheel model and the wheel model having the tire model calculates the longitudinal force and transfers it to the vehicle model.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) In the following, the present teaching is described in greater detail with reference to FIGS. 1 to 6 which, by way of example, show schematic and non-limiting advantageous embodiments of the present teaching. In the drawings:

(2) FIG. 1 shows an embodiment of a drive train test stand with a drive train as a test specimen,

(3) FIG. 2 shows an advantageous model structure of the simulation model,

(4) FIG. 3 shows an exemplary tire coordinate system,

(5) FIG. 4 shows a conventional speed control of a dynamometer on the drive train test stand,

(6) FIG. 5 shows a speed control according to the present teaching of a dynamometer on the drive train test stand, and

(7) FIG. 6a shows the dynamic state variables on a vehicle wheel and

(8) FIG. 6b shows the dynamic state variables on the dynamometer.

DETAILED DESCRIPTION

(9) FIG. 1 schematically shows a drive train test stand 1 for a drive train 2. The necessary, known test stand structures for transport, arrangement, storage and fixation of the drive train 2 and the other components on the drive train test stand 1 are not shown for reasons of clarity. In the exemplary embodiment shown, the drive train 2 comprises a drive unit 3, for example an internal combustion engine or an electric motor or a combination thereof, which is connected to a transmission 5 via a clutch 4 by means of a drive shaft. The transmission 5 is connected to a differential 6, via which in turn two side shafts 7a, 7b of the drive train 2 are driven. Dynamometers 8a, 8b, for example, electric motors, are arranged on the driven side shafts 7a, 7b of the drive train 2. The dynamometers 8a, 8b are arranged in a known manner and connected torque-proof to the side shafts 7a, 7b in order to be able to transmit a torque. For example, the dynamometers 8a, 8b are flanged to the wheel flange of the side shaft 7a, 7b. Of course, the drive train 2 could be implemented in the same way as a four-wheel drive train. In this case, dynamometers could be arranged in an analogous manner on all driven side shafts. Likewise, other drive concepts and designs of the drive train 2 are conceivable, such as a purely electrically driven axle (also in combination with a differently driven axle), or the use of wheel hub motors. Likewise, it does not matter how many axles or how many driven axles are present and how many wheels are arranged on an axle.

(10) The drive train 2 may also include control units, such as an engine control unit ECU or a transmission control unit TCU, in order to control the components of the drive train 2, in particular the drive unit 3, or optionally the drive units.

(11) The specific embodiment of the drive train 2 is irrelevant to the present teaching. All that is decisive is that at least one dynamometer 8a, 8b can be connected to at least one side shaft 7a, 7b. This is preferably a driven side shaft 7a, 7b, but it may also be a non-driven side shaft. The arrangement on a non-driven side shaft, for example, makes sense, if you also want to simulate a brake on a side shaft or the braking behavior.

(12) In addition, a test stand automation unit 10 is provided on the test stand 1, which controls the test run to be performed on the test stand 1. The test stand automation unit 10 is designed, for example, as a computer, or as a network of cooperating computers, with the required software. The test stand automation unit 10 controls, in particular, the dynamometers 8a, 8b, but also components of the drive train 2, in particular the drive unit 3 or the transmission 5. This can also be done via the control units ECU, TCU, for example by an accelerator pedal position being transmitted from the test stand automation unit 10 to the engine control unit ECU. The connection between the test stand automation unit 10 and the components of the drive train 2 or the drive train test stand 1 can also take place via a data bus 14, for example a conventional vehicle bus, as indicated in FIG. 1. However, the type of control of the components of the drive train 2 is irrelevant to the present teaching.

(13) In the test stand automation unit 10, a simulation model 20 is implemented, in the form of simulation hardware and/or simulation software, that simulates the movement of a virtual (i.e., simulated) vehicle with the drive train 2 through a virtual (i.e., simulated) test environment along a virtual route. The virtual test environment defines at least the virtual route (curves, gradients, road slopes, road surface). For this purpose, the route can be defined in advance. Often, real-world vehicles drive real distances and measure certain parameters (e.g., curves, gradients, road gradients, road surface (tire grip), vehicle speed, etc.). From such a real trip, a virtual route can then be generated. Likewise, a driving profile can be obtained from the real drive, so for example, the vehicle speed or a shifting action at certain points of the route or a speed change at certain points. The driving profile is implemented in the virtual test environment by a virtual driver, and for this purpose, different driver profiles can be defined, for example, a conservative or an aggressive driver, for example, who implement a desired change in speed differently or drive through a curve differently. Route and/or driving profile can also be defined completely freely by a user, for example in a suitable editor. However, the driving profile can also be generated, at least partially, at the test stand during the test run, in which an interface (e.g. steering wheel, accelerator pedal, clutch, brake pedal) is made available on the test stand via which a real user can control the virtual vehicle in the virtual test environment, for example, steer, step on the gas, brake, shift, etc. But the route can also be supplemented by events, such as traffic signs, traffic along the route, puddles or ice slabs on the roadway, etc.

(14) What the test run by the simulation of the virtual vehicle specifically looks like is irrelevant to the present teaching. It is only important that a movement of a virtual vehicle, in particular the statics, the dynamics, i.e., speeds and accelerations in space, and the interaction of a tire of the vehicle with the road are simulated in this movement. The simulation model 20 provides in predetermined time steps of the simulation, for example, with a frequency of 10 kHz, setpoint values for the drive unit 3, for example, a setpoint speed n.sub.A,set, a setpoint torque M.sub.A,set, a setpoint electric current or a setpoint electrical voltage set by the drive unit 3. Likewise, setpoint values for other components of the drive train 2 can be determined and transmitted, for example a shift command for the transmission 5. At the same time, the simulation model 20 supplies in the predetermined time steps of the simulation setpoint values for the dynamometers 8a, 8b used, preferably a setpoint speed n.sub.Ba,set, n.sub.Bb,set. In this way, the drive train 2 on the drive train test stand 1 experiences substantially the same conditions that the simulated vehicle would experience when driving along the simulated test track.

(15) The dynamometers 8a, 8b are controlled in a known manner in each case by an associated control unit 9a, 9b. For this purpose, a control unit 9a, 9b receives the setpoint value for the assigned dynamometer 8a, 8b, according to the present teaching a setpoint speed n.sub.Ba,set, n.sub.Bb,set, and controls this by means of the implemented controller, for example a known PI controller or a PID controller. Of course, the setpoint values for the individual dynamometers 8a, 8b do not need to be the same.

(16) To calculate the setpoint values and also to control the dynamometers 8a, 8b with the control units 9a, 9b, current values of the drive train 2 are measured on the drive train test stand 1, for example an actual speed n.sub.Ba,ist n.sub.Bb,ist of a dynamometer 8a, 8b with a speed measuring unit 15a, 15b and/or an actual torque M.sub.Ra,ist, Mb.sub.Rb,ist of a side shaft 7a, 7b with a torque measuring unit 16a, 16b, as indicated in FIG. 1. Instead of directly measuring these quantities, they could also be calculated by means of an observer from other measured quantities of the drive train 2. Likewise, these could be calculated from a model of the drive train 2. Likewise, other quantities could also be measured, calculated or estimated, such as a shaft torque M.sub.Ba, M.sub.Bb of the dynamometer 8a, 8b.

(17) To implement the simulation of the movement of the virtual vehicle, at least a vehicle model 22 which simulates the movement of the vehicle along the route, and a wheel model 21 with an integrated tire model 23 are required in the simulation model 20, as shown in FIG. 2. The wheel model 21 with the tire model 23 simulates the interaction of the wheel/tire with the environment, that is to say specifically with the road of the test track. In the tire model 23, the transmission of power from the tire 11 to the road is usually simulated and the wheel model 21 simulates the dynamics with the inertia of the vehicle wheel and with the forces/moments of the power transmission.

(18) For this purpose, a coordinate system based on the tire 11 can be used, as shown in FIG. 3. Here, a tire 11 of a vehicle wheel 19 is shown schematically on a generally curved roadway 12. The tire 11 stands on the roadway 12 at the wheel contact point P (FIG. 13 shows the tangential plane 13 on the curved roadway 12 in the wheel contact point P) and the tire 11 rotates about the wheel center C about a rotation axis y.sub.c. The tire 11 does not stand in a point P on the roadway 12, but on a tire contact area, which is commonly referred to as a contact patch L. The x-axis corresponds to the toe of the tire 11. The y-axis is the parallel of the rotational axis y.sub.c through the wheel contact point P and the z-axis is the connecting straight line through the wheel contact point P and the wheel center point C. The wheel contact point P is thus the point which minimizes the distance between the curved roadway 12 and the wheel center point C. This results in the tire 11 according to the selected coordinate system, a vertical force F.sub.z, a longitudinal force F.sub.x in the direction of the toe and a lateral force F.sub.y, a rolling resistance torque M.sub.y, a drilling torque M.sub.z and a tilting torque M.sub.x. These forces and torques are collectively referred to as tire force winder. The speed of the wheel contact point P of the vehicle observed in a roadway-fixed coordinate system is denoted by V(P). The projection of V(P) on the toe is called the longitudinal speed of the vehicle and is abbreviated to v.sub.x. A lateral velocity of the vehicle in the y-direction results in the same way.

(19) The wheel model 21 could then, for example, be implemented as an equation of motion of the form

(20) $J_W\stackrel{\mathrm{.Math.}}{}=M_y+\underset{M_{Fx}}{\underset{}{F_{x}r}}+M_R\left[+M_B+M_{\mathrm{aux}}\right]$
with the following variables:

(21) moment of inertia of the simulated vehicle wheel Jw, rotational acceleration {umlaut over ()} with the rotation angle (which can be measured, also as an analogous variable such as the speed), rolling resistance torque M.sub.y, longitudinal force F.sub.x, radius of the vehicle wheel r, a torque M.sub.R acting on the side shaft, for example, which is impressed by the drive unit 3 into the drive train 2, and other optional variables, such as a braking torque M.sub.B and any additional moments M.sub.aux, such as friction torques, drag torques, etc. The torques as algebraic quantities must be used with their correct algebraic signs. The mechanical connection between dynamometer 8a, 8b and side shaft 7a, 7b is generally regarded as, at least as sufficiently, stiff, so that the angle of rotation a can be derived in most cases from the measured actual speed n.sub.Ba,ist, n.sub.Bb,ist of the dynamometer 8a, 8b. The quantities brake torque M.sub.B and drive torque M.sub.R are either measured or are known from the test run, or are calculated or estimated in an observer from other variables measured on the drive train test stand 1. In any case, the wheel model 21 takes into account a torque M.sub.Fx as quantity of the tire 11 of the simulated vehicle wheel 19, which results from a longitudinal force F.sub.X, which is applied by the tire 11, preferably as an additional tire quantity also a rolling resistance torque M.sub.y.

(22) At least the longitudinal force F.sub.X is thereby calculated in the tire model 23 of the wheel model 21, but usually also at least the rolling resistance torque M.sub.y and often also the lateral force F.sub.y and the drilling torque M.sub.z, which tries to turn back the turned wheel. According to the current acting statics and dynamics (position, speed, acceleration) of the virtual vehicle, but also as a result of implemented drive concepts such as an active torque distribution, the acting tire forces and tire torque on the individual wheels of the vehicle of course do not need to be the same.

(23) Any known tire model 23 may be used to calculate the required tire forces and/or torques. Known tire models are for example a Pacejka model, a TameTire model, an Ftire model, a Delft-Tire model or an MF-SWIFT model. Frequently, the Pacejka model is used, for example, which is described in Pacejka H. B., et al., Tire Modeling for Use in Vehicle Dynamics Studies, International Congress and Exposition, Detroit, Feb. 23-27, 1987, SAE Technical Paper 870421. These tire models are well known and therefore will not be described in detail. In essence, a tire model calculates at least some of the aforementioned acting tire forces and/or tire torques. But which tire model is used is irrelevant to the present teaching. Likewise, in the wheel model 21 or in the tire model 23, the longitudinal slip and/or the transverse slip of the tire 11 can be taken into account, for example via a known relationship between the longitudinal force F.sub.X and lateral force F.sub.y and the longitudinal slip and transverse slip. This relationship can be stored, for example, in the form of a diagram or characteristic diagram, as described, for example, in EP 1 037 030 A2.

(24) The conventional known control of a dynamometer 8i on a drive train test stand 1, as described for example in DE 38 01 647 C2 or AT 508 031 B1, is shown in FIG. 4. The description is subsequently made only for one dynamometer 8i as a simulation for the i-th simulated vehicle wheel, but this also applies analogously to the other dynamometers used on the drive train test stand 1. In the vehicle model 22, for example, the longitudinal forces F.sub.Xi at the driven vehicle wheels and from other known or parameterized geometrical or kinematic quantities, such as the vehicle mass, the road gradient, the road inclination, the slip angle of the i-th wheel, the currently acting slip, the air resistance, the pitching and rolling motion of the vehicle, the road condition, etc., the state variables of the vehicle, in particular the longitudinal speed v.sub.x of the vehicle and in most cases also the yaw rate are calculated. In the vehicle model 22, the current vertical forces F.sub.zi are also calculated as tire contact forces on the vehicle wheels from the current acting statics (weight force) and dynamics (position, speed, acceleration in space) of the simulated vehicle. From this, along with a coefficient of adhesion between road and tire, the longitudinal force F.sub.X on a wheel can be calculated, for example, in the tire model 23 of the wheel model 21. In the wheel model 21i of the i-th vehicle wheel, or in the tire model 23i integrated therein, a rolling resistance torque M.sub.yi of the i-th vehicle wheel may also be calculated. For this purpose, the actual torque M.sub.Ri,ist is determined (measured (torque measuring unit 16i, estimated) on the side shaft 7i and provided to the wheel model 21i. At least from the longitudinal velocity v.sub.x of the simulated vehicle and the vertical force F.sub.zi, and optionally an actual speed n.sub.Bi,act of the i-th vehicle wheel, and the actual torque M.sub.Ri,ist on the associated side shaft 7i, and optionally also with the rolling torque M.sub.yi and other torques, the longitudinal force F.sub.Xi is calculated, which in turn is provided to the vehicle model 22. For this purpose, of course, required parameters, such as a current coefficient of adhesion between the route and tire 11 or a slip angle are also available. Likewise, in the calculation of the longitudinal force F.sub.Xi, in particular for realistic and also highly dynamic test runs, the slip can also be taken into account. From the longitudinal velocity v.sub.x calculated in the current time step of the simulation, the setpoint rotational speed n.sub.Bi,set to be set the side shaft 7 by the dynamometer 8i is calculated in the wheel model 21i, which is transmitted to the associated control unit 9i with an implemented rotational speed controller 17i as the setpoint value of the control. The speed controller 17i calculates therefrom a control torque M.sub.REi according to the implemented controller, which is transferred to the dynamometer 8i as torque M.sub.Bi,soll to be set. This manipulated variable for the dynamometer 8i is converted in the power electronics (e.g. an inverter) of the dynamometer 8i in a known manner into an electric motor current. The speed controller 17i is designed, for example, as a conventional feedback controller which regulates a control error as a deviation between an actual speed n.sub.Bi,ist (measured (speed measuring unit 15i, for example) of the side shaft 7i and the setpoint speed n.sub.Bi,set.

(25) The speed control of a dynamometer 8i according to the present teaching will now be explained with reference to FIG. 5. Compared to the conventional control as described with FIG. 4, the speed control of the dynamometer 8i is supplemented according to the present teaching by a moment of inertia compensation. The moment of inertia compensation compensates for the deviation between the known moment of inertia J.sub.Bi of the dynamometer 8i and the known moment of inertia j.sub.Ri of the virtual simulated vehicle wheel. In other words, this takes into account the deviation between the simulation and the reality on the drive train test stand 1 and reduces its influence.

(26) For this purpose, the wheel model 21i also transfers at the interface 30 of the test stand automation unit 10 a torque M.sub.Fxi, which is caused by the longitudinal force F.sub.Xi, to the control unit 9i, in addition to the setpoint speed n.sub.Bi,set, as in the prior art. Equivalently, of course, the longitudinal force F.sub.Xi, possibly also with the radius of the vehicle wheel r.sub.i, could be transferred. This is therefore also understood in the context of the present teaching as transferring the torque M.sub.Fxi. Since this torque M.sub.Fxi has to be calculated anyway in the wheel model 21i for carrying out the test run, it is not necessary to adapt the wheel model 21i. Merely the transfer of an additional variable between the simulation model 20 and the control unit 9i is to be provided, for example, an additional interface to the test stand automation unit 10 for transferring the torque M.sub.Fxi, which is easy to implement.

(27) In the control unit 9i, a compensation unit 18i is provided which calculates a compensation torque M.sub.Ki from the torque M.sub.Fxi and a deviation A.sub.Ji between a moment of inertia J.sub.Bi of the dynamometer and a moment of inertia J.sub.Ri of the simulated vehicle wheel, i.e. M.sub.Ki=f(M.sub.Fxi,A.sub.Ji). The calculated compensation torque M.sub.Ki is preferably recalculated in each time step of the control. This compensation torque M.sub.Ki is added to the torque M.sub.REi, which is calculated in the speed controller 17i of the control unit 9i according to the implemented control law (e.g., a conventional PI or PID controller). This sum torque is then specified to the dynamometer 8i as torque M.sub.Bi,soll.

(28) To determine the deviation A.sub.Ji for the i-th vehicle wheel 19i, the procedure may be as follows, reference being made to FIG. 6. On a vehicle wheel 19i with moment of inertia J.sub.Ri, which rotates at the angular speed .sub.Ri, the torque M.sub.Fxi, which is caused by the longitudinal force F.sub.Xi, and the torque acting on the side shaft 7i M.sub.Ri (FIG. 6a) are acting. According to the principle of angular momentum, one gets J.sub.Ri.sub.Ri=M.sub.RiM.sub.Fxi. On the shaft of the dynamometer 8i, which rotates at the angular speed .sub.Bi, a loading torque M.sub.Bi (FIG. 1) acts and the dynamometer 8i applies the torque M.sub.Di (FIG. 6b). According to the principle of angular momentum, one gets J.sub.Bi{dot over ()}.sub.Bi=M.sub.BiM.sub.Di. The dynamometer 8i should now emulate the vehicle wheel 19i as well as possible on the drive train test stand 1. It can therefore be demanded that the rotational accelerations on the side shaft 7i and on the shaft of the dynamometer 8i are the same. i.e. {dot over ()}.sub.Ri={dot over ()}.sub.Bi, from which it is possible to derive directly from the two principle of angular momentum equations

(29) $M_{\mathrm{Di}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+M_{\mathrm{Ri}}-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Bi}}.$

(30) Assuming an (at least sufficiently) high mechanical rigidity of the mechanical connection of the dynamometer 8i to the drive train 2, the torque M.sub.Ri acting on the side shaft 7i can be equated with the torque M.sub.Bi of the dynamometer 8i for the sake of simplicity, which leads to

(31) $M_{\mathrm{Di}}=\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}{M}_{\mathrm{Fxi}}+\left(1-\frac{J_{\mathrm{Bi}}}{J_{\mathrm{Ri}}}\right)M_{\mathrm{Ri}}.$

(32) Therefore, only one of the two torques has to be determined on the drive train test stand 1. Thus, the deviation A.sub.Ji is realized as quotient of the moment of inertia J.sub.Bi of the dynamometer 8i and the moment of inertia J.sub.Ri of the simulated vehicle wheel 19i. Alternatively one can also demand M.sub.Bi=M.sub.Ri, from which one can derive directly from the two principle of angular momentum equations
M.sub.Di=M.sub.Fxi+J.sub.Ri{dot over ()}.sub.RiJ.sub.Bi{dot over ()}.sub.Bi.

(33) Assuming an (at least sufficiently) high mechanical rigidity of the mechanical connection of the dynamometer 8i to the drive train 2, the speed n.sub.Ri (or .sub.Ri) on the side shaft 7i can be equated with the speed n.sub.Bi (or .sub.Bi) of the dynamometer 8i for the sake of simplicity, which leads to
M.sub.Di=M.sub.Fxi+(J.sub.RiJ.sub.Bi){dot over ()}.sub.Ri.

(34) Therefore only one of the two speeds has to be determined on the drive train test stand 1. Thus, the deviation A.sub.ji is realized as the difference between the moment of inertia J.sub.Ri of the simulated vehicle wheel 19i and the moment of inertia J.sub.Bi of the load machine 8i. If the inertia moments J.sub.Bi, J.sub.Ri are equal, the equations are reduced to M.sub.Di=M.sub.Fxi. The torques M.sub.Bi and/or M.sub.Ri, or the speeds n.sub.Bi and/or n.sub.Ri can in turn be measured, calculated or estimated and can therefore be assumed to be known for the moment of inertia compensation.

(35) So that the dynamometer 8i emulates the vehicle wheel 19i well despite different moments of inertia J.sub.Ri, J.sub.Bi, the dynamometer 8i would therefore have to apply the torque M.sub.Di. The compensation torque M.sub.Ki is therefore set equal to this torque M.sub.Di. The compensation torque M.sub.Ki can therefore also be seen as a master manipulated variable, with the speed controller 17i then only having to compensate for any deviations. The requirements of the speed controller 17i, for example of the gain, can thus also be reduced and at the same time the dynamics of the speed controller 17i (in terms of a rate of change of the manipulated variable) and the speed can be improved. Likewise, this can increase the stability of the speed controller 17i.

(36) Of course, other tire quantities, in particular a rolling resistance torque M.sub.yi, could also be taken into account in the above-described principles of angular momentum equations for the vehicle wheel 19i. This would require further interfaces between the simulation model 20 and the control unit 9.

(37) For the present teaching, it is irrelevant how the drive train 2 is arranged on the drive train test stand 1. The whole real vehicle could also be arranged on the drive train test stand 1 and only the vehicle wheels, at least the driven ones, could be replaced by dynamometers 8i. Likewise, the real vehicle with vehicle wheels could be arranged on a roller on the drive train test stand 1. The dynamometer 8i would drive the roller and thus act indirectly on the drive train 2 and the dynamometer 8i would thereby be indirectly connected to a side shaft. Also, a plurality of rollers could be provided, for example, one roller per driven vehicle wheel or per axle. With such a roller test stand, however, it would then normally not be possible to carry out test runs with high dynamics.